WO2022224446A1 - 画像処理装置、画像処理方法及び記憶媒体 - Google Patents
画像処理装置、画像処理方法及び記憶媒体 Download PDFInfo
- Publication number
- WO2022224446A1 WO2022224446A1 PCT/JP2021/016473 JP2021016473W WO2022224446A1 WO 2022224446 A1 WO2022224446 A1 WO 2022224446A1 JP 2021016473 W JP2021016473 W JP 2021016473W WO 2022224446 A1 WO2022224446 A1 WO 2022224446A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- score
- classification
- image processing
- image
- unit
- Prior art date
Links
- 238000012545 processing Methods 0.000 title claims abstract description 119
- 238000003672 processing method Methods 0.000 title claims description 8
- 238000004364 calculation method Methods 0.000 claims abstract description 107
- 238000003384 imaging method Methods 0.000 claims description 19
- 230000003902 lesion Effects 0.000 description 69
- 238000000034 method Methods 0.000 description 26
- 230000008569 process Effects 0.000 description 18
- 230000015654 memory Effects 0.000 description 17
- 238000001839 endoscopy Methods 0.000 description 15
- 238000012986 modification Methods 0.000 description 10
- 230000004048 modification Effects 0.000 description 10
- 238000001514 detection method Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 210000001072 colon Anatomy 0.000 description 4
- 238000013459 approach Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 208000023514 Barrett esophagus Diseases 0.000 description 2
- 208000023665 Barrett oesophagus Diseases 0.000 description 2
- 206010025323 Lymphomas Diseases 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 2
- 206010028980 Neoplasm Diseases 0.000 description 2
- 230000002183 duodenal effect Effects 0.000 description 2
- 210000003238 esophagus Anatomy 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 208000027866 inflammatory disease Diseases 0.000 description 2
- 230000001613 neoplastic effect Effects 0.000 description 2
- 210000000056 organ Anatomy 0.000 description 2
- 238000001617 sequential probability ratio test Methods 0.000 description 2
- 210000000813 small intestine Anatomy 0.000 description 2
- 210000002784 stomach Anatomy 0.000 description 2
- 206010009944 Colon cancer Diseases 0.000 description 1
- 208000011231 Crohn disease Diseases 0.000 description 1
- 206010061825 Duodenal neoplasm Diseases 0.000 description 1
- 208000000289 Esophageal Achalasia Diseases 0.000 description 1
- 208000000461 Esophageal Neoplasms Diseases 0.000 description 1
- 208000000624 Esophageal and Gastric Varices Diseases 0.000 description 1
- 206010061968 Gastric neoplasm Diseases 0.000 description 1
- 208000007882 Gastritis Diseases 0.000 description 1
- 201000003741 Gastrointestinal carcinoma Diseases 0.000 description 1
- 208000034991 Hiatal Hernia Diseases 0.000 description 1
- 206010020028 Hiatus hernia Diseases 0.000 description 1
- 206010061218 Inflammation Diseases 0.000 description 1
- 206010030136 Oesophageal achalasia Diseases 0.000 description 1
- 206010030155 Oesophageal carcinoma Diseases 0.000 description 1
- 206010030216 Oesophagitis Diseases 0.000 description 1
- 208000009565 Pharyngeal Neoplasms Diseases 0.000 description 1
- 206010034811 Pharyngeal cancer Diseases 0.000 description 1
- 206010039580 Scar Diseases 0.000 description 1
- 206010054184 Small intestine carcinoma Diseases 0.000 description 1
- 208000005718 Stomach Neoplasms Diseases 0.000 description 1
- 208000007107 Stomach Ulcer Diseases 0.000 description 1
- 206010056091 Varices oesophageal Diseases 0.000 description 1
- 201000000621 achalasia Diseases 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 239000002775 capsule Substances 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 206010009887 colitis Diseases 0.000 description 1
- 208000029742 colonic neoplasm Diseases 0.000 description 1
- 238000002052 colonoscopy Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 208000000718 duodenal ulcer Diseases 0.000 description 1
- 206010013864 duodenitis Diseases 0.000 description 1
- 210000001198 duodenum Anatomy 0.000 description 1
- 201000000312 duodenum cancer Diseases 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 201000004101 esophageal cancer Diseases 0.000 description 1
- 208000024170 esophageal varices Diseases 0.000 description 1
- 201000010120 esophageal varix Diseases 0.000 description 1
- 208000006881 esophagitis Diseases 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 206010017758 gastric cancer Diseases 0.000 description 1
- 230000002496 gastric effect Effects 0.000 description 1
- 201000005917 gastric ulcer Diseases 0.000 description 1
- 208000014617 hemorrhoid Diseases 0.000 description 1
- 230000004054 inflammatory process Effects 0.000 description 1
- 201000002313 intestinal cancer Diseases 0.000 description 1
- 230000000968 intestinal effect Effects 0.000 description 1
- 201000008267 intestinal tuberculosis Diseases 0.000 description 1
- 210000002429 large intestine Anatomy 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 208000003154 papilloma Diseases 0.000 description 1
- 208000014081 polyp of colon Diseases 0.000 description 1
- 208000015768 polyposis Diseases 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 201000011549 stomach cancer Diseases 0.000 description 1
- 208000023984 stomach polyp Diseases 0.000 description 1
- 210000000115 thoracic cavity Anatomy 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 208000018408 tumor of duodenum Diseases 0.000 description 1
- 208000019553 vascular disease Diseases 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 230000003936 working memory Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
- A61B1/00002—Operational features of endoscopes
- A61B1/00004—Operational features of endoscopes characterised by electronic signal processing
- A61B1/00009—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope
- A61B1/000094—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope extracting biological structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/17—Image acquisition using hand-held instruments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/62—Extraction of image or video features relating to a temporal dimension, e.g. time-based feature extraction; Pattern tracking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
- A61B1/00002—Operational features of endoscopes
- A61B1/00004—Operational features of endoscopes characterised by electronic signal processing
- A61B1/00009—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope
- A61B1/000096—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope using artificial intelligence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10068—Endoscopic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Definitions
- the present disclosure relates to the technical field of image processing apparatuses, image processing methods, and storage media that process images acquired in endoscopy.
- Patent Literature 1 discloses a method of learning a learning model that outputs information about a lesion site included in captured image data when captured image data generated by an imaging device is input.
- Patent Literature 2 discloses a classification method for classifying series data by a method applying a sequential probability ratio test (SPRT).
- SPRT sequential probability ratio test
- one object of the present disclosure is to provide an image processing apparatus, an image processing method, and a storage medium capable of appropriately detecting an attention point such as a lesion site in endoscopy. is.
- Acquisition means for acquiring a photographed image obtained by photographing an object to be inspected by an imaging unit provided in the endoscope; score calculation means for calculating a score relating to the likelihood that a spot of interest exists in the photographed image based on the plurality of photographed images in time series; Classification means for classifying the plurality of captured images based on the scores; It is an image processing apparatus having
- the computer Acquiring a photographed image of an object to be inspected by an imaging unit provided in the endoscope, calculating a score related to the likelihood that a spot of interest exists in the captured image based on a plurality of the captured images in time series; classifying the plurality of captured images based on the score; It is an image processing method.
- One aspect of the storage medium is Acquiring a photographed image of an object to be inspected by an imaging unit provided in the endoscope, calculating a score related to the likelihood that a spot of interest exists in the captured image based on a plurality of the captured images in time series;
- a storage medium storing a program that causes a computer to execute processing for classifying the plurality of captured images based on the scores.
- FIG. 1 shows a schematic configuration of an endoscopy system
- 2 shows the hardware configuration of an image processing apparatus
- 1 is a functional block diagram of an image processing device
- FIG. 4 is a graph showing changes in classification scores calculated based on the first mode
- 9 is a graph showing changes in classification scores calculated based on the second mode.
- 1 shows a first display example of a display screen displayed by a display device in an endoscopy.
- a second display example of a display screen displayed by the display device 2 during an endoscopy is shown.
- a third display example of a display screen displayed by the display device 2 during an endoscopy is shown. It is an example of the flowchart which showed the procedure of the process based on a 1st aspect.
- FIG. 11 is a block diagram of an image processing apparatus according to a second embodiment; FIG. It is an example of a flowchart executed by an image processing apparatus in the second embodiment.
- FIG. 1 shows a schematic configuration of an endoscopy system 100.
- an endoscopy system 100 is a system that presents a site suspected of having a lesion (lesion site) to an examiner such as a doctor who performs an examination or treatment using an endoscope. It mainly includes an image processing device 1 , a display device 2 , and an endoscope 3 connected to the image processing device 1 .
- the image processing device 1 acquires images (also referred to as “captured images Ia”) captured by the endoscope 3 in time series from the endoscope 3 and displays a screen based on the captured images Ia on the display device 2 .
- the captured image Ia is an image captured at a predetermined frame cycle during at least one of the process of inserting the endoscope 3 into the subject and the process of removing the endoscope 3 .
- the image processing apparatus 1 analyzes the captured image Ia to detect the presence of a lesion site from the captured image Ia, and causes the display device 2 to display information about the detection result.
- the display device 2 is a display or the like that performs a predetermined display based on a display signal supplied from the image processing device 1 .
- the endoscope 3 mainly includes an operation unit 36 for an examiner to perform predetermined input, a flexible shaft 37 inserted into an organ to be imaged by an examinee, and an ultra-compact imaging device. It has a tip portion 38 containing an imaging unit such as an element, and a connection portion 39 for connecting to the image processing apparatus 1 .
- the configuration of the endoscopy system 100 shown in FIG. 1 is an example, and various modifications may be made.
- the image processing device 1 may be configured integrally with the display device 2 .
- the image processing device 1 may be composed of multiple devices.
- endoscopes targeted in the present disclosure include, for example, pharyngeal endoscopes, bronchoscopes, upper gastrointestinal endoscopes, duodenal endoscopes, small intestine endoscopes, colonoscopes, capsule endoscopes, thoracic scopes, laparoscopes, cystoscopes, choledoscopes, arthroscopes, spinal endoscopes, angioscopes, epidural endoscopes, and the like.
- the disease conditions of the lesion site to be detected in the present disclosure are exemplified as (a) to (f) below.
- Esophagus esophageal cancer, esophagitis, hiatal hernia, Barrett's esophagus, esophageal varices, esophageal achalasia, esophageal submucosal tumor, esophageal benign tumor
- Stomach gastric cancer, gastritis, gastric ulcer, gastric polyp, gastric tumor
- Duodenum duodenal cancer, duodenal ulcer, duodenitis, duodenal tumor, duodenal lymphoma
- Small intestine small bowel cancer, small bowel neoplastic disease, small bowel inflammatory disease , small intestinal vascular disease
- colon colon cancer
- FIG. 2 shows the hardware configuration of the image processing apparatus 1.
- the image processing device 1 mainly includes a processor 11 , a memory 12 , an interface 13 , an input section 14 , a light source section 15 and a sound output section 16 . Each of these elements is connected via a data bus 19 .
- the processor 11 executes a predetermined process by executing a program or the like stored in the memory 12.
- the processor 11 is a processor such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a TPU (Tensor Processing Unit).
- Processor 11 may be composed of a plurality of processors.
- Processor 11 is an example of a computer.
- the memory 12 is composed of various volatile memories used as working memory, such as RAM (Random Access Memory) and ROM (Read Only Memory), and non-volatile memory for storing information necessary for processing of the image processing apparatus 1. be done.
- the memory 12 may include an external storage device such as a hard disk connected to or built into the image processing apparatus 1, or may include a storage medium such as a detachable flash memory.
- the memory 12 stores a program for the image processing apparatus 1 to execute each process in this embodiment.
- the memory 12 functionally has a first calculation information storage unit D1 that stores first calculation information and a second calculation information storage unit D2 that stores second calculation information.
- the first calculation information and the second calculation information are information used when the image processing apparatus 1 classifies the presence or absence of a lesion site or information indicating the result of calculation regarding the classification, and the details will be described later.
- the memory 12 stores various parameters necessary for calculating a classification score regarding the presence or absence of a lesion site. At least part of the information stored in the memory 12 may be stored by an external device other than the image processing device 1 .
- the above external device may be one or a plurality of server devices capable of data communication with the image processing device 1 via a communication network or the like or by direct communication.
- the interface 13 performs an interface operation between the image processing device 1 and an external device.
- interface 13 provides display information “Ib” generated by processor 11 to display device 2 .
- the interface 13 also supplies the endoscope 3 with light or the like generated by the light source unit 15 .
- the interface 13 also supplies the processor 11 with an electrical signal indicating the captured image Ia supplied from the endoscope 3 .
- the interface 13 may be a communication interface such as a network adapter for performing wired or wireless communication with an external device, and may be a hardware interface conforming to USB (Universal Serial Bus), SATA (Serial AT Attachment), or the like. may
- the input unit 14 generates an input signal based on the operation by the inspector.
- the input unit 14 is, for example, a button, touch panel, remote controller, voice input device, or the like.
- the light source unit 15 generates light to be supplied to the distal end portion 38 of the endoscope 3 .
- the light source unit 15 may also incorporate a pump or the like for sending out water or air to be supplied to the endoscope 3 .
- the sound output unit 16 outputs sound under the control of the processor 11 .
- lesion portion detection processing by the image processing apparatus 1 will be described.
- the image processing apparatus 1 detects a lesion site based on a variable number of time-series photographed images Ia. As a result, the image processing apparatus 1 accurately detects a lesion site that is difficult to distinguish from a single image, and presents the detection result.
- Lesion sites that are difficult to distinguish from a single image include, for example, non-cancerous portions and flat cancerous portions in Barrett's esophagus.
- FIG. 3 is a functional block diagram of the image processing device 1.
- the processor 11 of the image processing apparatus 1 functionally includes a photographed image acquisition section 30 , a score calculation section 31 , a classification section 34 and a display control section 35 .
- the blocks that exchange data are connected by solid lines, but the combinations of blocks that exchange data are not limited to those shown in FIG. The same applies to other functional block diagrams to be described later.
- the captured image acquisition unit 30 acquires captured images Ia captured by the endoscope 3 via the interface 13 at predetermined intervals according to the frame cycle of the endoscope 3 . Then, the captured image acquisition unit 30 supplies the acquired captured image Ia to the score calculation unit 31 and the display control unit 35, respectively.
- the score calculation unit 31 is used for classification regarding the presence or absence of a lesion site, and calculates a score (also referred to as a "classification score") that indicates the likelihood that a lesion site exists.
- the score calculator 31 determines a classification score based on the likelihood ratio for the time-series captured images Ia.
- the “likelihood ratio for the time-series photographed images Ia” is the ratio of the likelihood that a lesion site exists in the time-series photographed images Ia to the likelihood that a lesion site does not exist in the time-series photographed images Ia. Point.
- the likelihood ratio increases as the likelihood of existence of a lesion site increases.
- the score calculator 31 functionally has a first calculator 32 and a second calculator 33 .
- the first calculation unit 32 calculates the likelihood ratio for the latest “N” (N is an integer) captured images Ia at each processing time for each predetermined length of time, and sends the calculation result to the first calculation unit 32. supply.
- the first calculation unit 32 for example, when N photographic images Ia are input, the likelihood ratio calculation learned to output the likelihood ratio regarding the input N photographic images Ia Calculate the likelihood ratio using the model.
- the likelihood ratio calculation model may be a deep learning model, any other machine learning model, or a statistical model.
- the memory 12 stores learned parameters of the likelihood ratio calculation model, and the first calculation unit 32 updates the likelihood ratio calculation model configured by referring to the parameters. N captured images Ia are input, and the likelihood ratio output by the model is obtained.
- the likelihood ratio calculation model is composed of a neural network, for example, various parameters such as layer structure, neuron structure of each layer, the number and size of filters in each layer, and the weight of each element of each filter are stored in the memory 12 in advance. It is
- the likelihood ratio calculation model may include any feature amount extractor that extracts the feature amount (that is, feature vector) of each captured image Ia input to the likelihood ratio calculation model. It may be configured separately from the feature amount extractor. In the latter case, the likelihood ratio calculation model outputs the likelihood ratio for the N photographed images Ia when the feature quantity of the N photographed images Ia extracted by the feature quantity extractor is input. It is a model that has been trained to Also, the feature quantity extractor preferably extracts a feature quantity representing the relationship of time series data based on any method for calculating the relationship of time series data such as LSTM (Long Short Term Memory). may be
- the first calculation unit 32 stores the likelihood ratio calculated at each processing time and the data used by the first calculation unit 32 to calculate the likelihood ratio as first calculation information in the first calculation information storage unit D1.
- the “data used to calculate the likelihood ratio” may be the captured images Ia used to calculate the likelihood ratio, or may be feature amounts extracted from these captured images Ia.
- the second calculation unit 33 calculates a likelihood ratio (also referred to as an “integrated likelihood ratio”) by integrating the likelihood ratios calculated at the current processing time and the past processing time, and classifies based on the integrated likelihood ratio. determine the score.
- a likelihood ratio also referred to as an “integrated likelihood ratio”
- the classification score may be the integrated likelihood ratio itself, or may be a function including the integrated likelihood ratio as a variable.
- Equation (1) represents the probability of belonging to each class (that is, confidence between 0 and 1).
- the likelihood ratio stored in the first calculation information storage unit D1 by the first calculation unit 32 as the first calculation information can be used.
- Equation (1) the current processing time t increases with the passage of time, so the time-series length of the captured image Ia (or its feature amount) used to calculate the integrated likelihood ratio is variable.
- the second calculation unit 33 can calculate the classification score in consideration of the variable number of captured images Ia as a first advantage.
- the second advantage is that it is possible to classify time-dependent features
- the third advantage is that classification is possible even with difficult-to-discriminate data with little loss of accuracy.
- a score can be preferably calculated.
- the second calculation unit 33 stores the integrated likelihood ratio and the classification score calculated at each processing time as second calculation information in the second calculation information storage unit D2.
- the classification unit 34 Based on the classification scores calculated by the second calculation unit 33, the classification unit 34 performs classification regarding the presence or absence of lesion sites. In this case, the classification unit 34 performs the above-described classification using a classification score threshold regarding the presence or absence of a lesion site. In the present embodiment, as an example, the higher the classification score, the higher the possibility that a lesion site is present. ) In the above cases, it is determined that a lesion site exists in the photographed image Ia used to calculate the classification score. On the other hand, if the classification score is equal to or less than a predetermined threshold lower than the first threshold (also referred to as a “second threshold Th2”), the classification unit 34 determines that the photographed image Ia used to calculate the classification score contains a lesion site.
- a predetermined threshold lower than the first threshold also referred to as a “second threshold Th2”
- the first threshold Th1 and the second threshold Th2 are, for example, adapted values determined based on experiments and the like, and are stored in advance in the memory 12 or the like. Then, while the classification score does not reach these thresholds, the classification unit 34 does not determine the classification, and suspends (undecides) the classification of the target captured image Ia.
- “to classify” includes not only determining (confirming) the classification but also suspending the classification. A specific mode of classification by the classification unit 34 will be described later.
- the classification unit 34 supplies the classification result to the display control unit 35 .
- the classification section 34 notifies the score calculation section 31 that the classification has been determined.
- the first calculation unit 32 and the second calculation unit 33 of the score calculation unit 31 reset (initialize) the processing result before the notification, and Calculate the classification score based on A specific example of this processing will be described in detail with reference to FIGS. 4 and 5.
- FIG. 4 A specific example of this processing will be described in detail with reference to FIGS. 4 and 5.
- the display control unit 35 generates the display information Ib based on the captured image Ia and the classification result supplied from the classification unit 34, and supplies the display information Ib to the display device 2 via the interface 13.
- the display device 2 is caused to display the image Ia and the information about the classification result by the classification unit 34 .
- the display control unit 35 may cause the display device 2 to further display the information about the classification scores stored in the second calculation information storage unit D2. A display example of the display control unit 35 will be described later.
- each component of the captured image acquisition unit 30, the score calculation unit 31, the classification unit 34, and the display control unit 35 can be realized by the processor 11 executing a program, for example. Further, each component may be realized by recording necessary programs in an arbitrary nonvolatile storage medium and installing them as necessary. Note that at least part of each of these components may be realized by any combination of hardware, firmware, and software, without being limited to being implemented by program software. Also, at least part of each of these components may be implemented using a user-programmable integrated circuit, such as an FPGA (Field-Programmable Gate Array) or a microcontroller. In this case, this integrated circuit may be used to implement a program composed of the above components.
- FPGA Field-Programmable Gate Array
- each component may be configured by an ASSP (Application Specific Standard Produce), an ASIC (Application Specific Integrated Circuit), or a quantum processor (quantum computer control chip).
- ASSP Application Specific Standard Produce
- ASIC Application Specific Integrated Circuit
- quantum processor quantum computer control chip
- the score calculation unit 31 calculates the start point of the captured image Ia used for calculating the classification score. (also called the "starting point") to initialize the classification score.
- the score calculator 31 sets a start point at predetermined time intervals, and calculates classification scores in parallel for each set start point.
- FIG. 4 shows graphs "G1" to "G4" showing changes in classification scores calculated based on the first mode.
- the score calculator 31 starts processing at time “t0” and calculates a classification score based on the captured image Ia obtained in the frame cycle of the endoscope 3 .
- the classification unit 34 may supply the display control unit 35 with information indicating that the classification is undecided.
- the classification unit 34 supplies the display control unit 35 with a classification result indicating that the photographed images Ia obtained from the time t0 to the time t1 do not include the lesion site, and the classification score A notification for resetting the calculation process (that is, resetting the starting point) is supplied to the score calculation unit 31 .
- the score calculation unit 31 resets the classification score calculation process based on the notification from the classification unit 34 that the classification has been determined. Then, the score calculation unit 31 starts the classification score calculation processing for the photographed images Ia obtained after the time t1. After that, as shown in the graph G2, at time "t2", the classification score calculated by the score calculation unit 31 based on the captured images Ia obtained from time t1 to time t2 becomes equal to or less than the second threshold Th2. Therefore, in this case, the classification unit 34 generates a classification result indicating that the photographed image Ia obtained from time t1 to time t2 does not include a lesion site. Then, the classification unit 34 supplies the generated classification result to the display control unit 35 and supplies a notification for resetting the classification score calculation process to the score calculation unit 31 .
- the score calculation unit 31 starts classification score calculation processing for the captured images Ia obtained after time t2, based on the notification from the classification unit 34 that the classification has been performed.
- the photographed image Ia obtained from time t2 to time t3 includes a lesion site, as shown in graph G3, the classification score calculated by the score calculation unit 31 after time t2 is , gradually approaches the first threshold Th1.
- the classification score calculated by the score calculation unit 31 based on the captured images Ia obtained from time t2 to time “t3” becomes equal to or greater than the first threshold Th1. Therefore, in this case, the classification unit 34 generates a classification result that the photographed image Ia obtained from the time t2 to the time t3 includes the lesion site. Then, the classification unit 34 supplies the generated classification result to the display control unit 35 and supplies a notification for resetting the classification score calculation process to the score calculation unit 31 . Then, at time t3, the score calculation unit 31 starts classification score calculation processing for the captured image Ia obtained after time t2 based on the notification that the classification has been determined from the classification unit 34 ( See graph G4).
- the score calculation unit 31 considers that the imaging target of the endoscope 3 changes over time, and the classification score is equal to or greater than the first threshold Th1 or less than the second threshold Th2. (ie, when a classification determination result is obtained), the classification score calculation process is reset. As a result, the score calculation unit 31 preferably eliminates the influence of the determination result of the past imaging target site, and calculates the classification score so that the classification score appropriately reaches the first threshold Th1 when the lesion site is the subject. can be calculated.
- FIG. 5 is graphs "G10" to "G17” showing changes in classification scores calculated based on the second aspect.
- the score calculator 31 starts processing at time “t10” and calculates a classification score based on the captured image Ia obtained in a predetermined frame cycle.
- the lesion site is included in the photographed image Ia in the vicinity of the period from time "t14" to time "t16".
- the score calculation unit 31 sets the start points of the captured images Ia used for calculating the classification score at predetermined time intervals from time t10, and for each start point, the captured images Ia generated after the target start point.
- the "predetermined time interval" is, for example, an interval that is an integral multiple of the frame cycle, and more specifically, a time interval that corresponds to the processing capability of the image processing apparatus 1 (for example, a time interval that enables real-time processing). ).
- Each thread continues to calculate the classification score until it reaches either the first threshold Th1 or the second threshold Th2.
- calculation of classification scores with different starting points is performed in parallel. For example, in the period from time t11 to time t12, there are a thread whose start point is time t10 (see graph G10) and a thread whose start point is time t11 (see graph G11).
- the score calculation unit 31 calculates the classification score calculated at predetermined time intervals, so that when the lesion site is the subject, the classification score appropriately reaches the first threshold value Th1.
- a classification score can be calculated that achieves As a result, the classification unit 34 can appropriately classify a group of photographed images Ia in which a lesion site is a subject as an image group in which a lesion site is present.
- FIG. 6 shows a first display example of a display screen displayed by the display device 2 during an endoscopy.
- the display control unit 35 of the image processing device 1 outputs the display information Ib generated based on the captured image Ia acquired by the captured image acquisition unit 30 and the classification result generated by the classification unit 34 to the display device 2 .
- the display control unit 35 causes the display device 2 to display the display screen shown in FIG. 6 by transmitting the captured image Ia and the display information Ib to the display device 2 .
- the display control unit 35 of the image processing apparatus 1 provides the latest image display area 70 representing the latest captured image Ia acquired by the captured image acquiring unit 30, and displays information on the lesion site on the display screen.
- the score calculation unit 31 calculates the classification score based on the first mode, and the latest classification score calculated by the score calculation unit 31 corresponds to either the first threshold Th1 or the second threshold Th2. has not yet reached As a result, the classification unit 34 cannot determine the classification. Therefore, the display control unit 35 recognizes that the classification unit 34 has not been able to determine the classification from the classification score or the notification from the classification unit 34, and the classification regarding the presence or absence of the lesion site is undetermined (that is, the lesion site exists). is being determined) is displayed on the display screen. In addition to the information described above, the display control unit 35 also displays the latest classification score (-0.5 in this case) and the description of the threshold on the display screen.
- the display control unit 35 when the classification of the captured image Ia is undecided, the display control unit 35 outputs information indicating that the classification is undecided together with information about the classification score. Thereby, the display control unit 35 can preferably visualize the current state regarding the presence/absence determination of the lesion site.
- the display control unit 35 may display the graph shown in FIG. 4 on the display screen together with or instead of displaying the classification scores. Further, when the score calculation unit 31 calculates the classification score based on the second mode, the display control unit 35 may display each classification score corresponding to each of the active threads, The classification score closest to either Th1 or the second threshold Th2 may be displayed, and the graph shown in FIG. 5 may be displayed on the display screen. Alternatively, or in addition to the display control shown in the first display example, if the classification of the captured image Ia is undecided, the display control unit 35 provides voice guidance or The sound output unit 16 may be instructed to output a predetermined warning sound. This also allows the display control unit 35 to preferably let the examiner know that the classification is undetermined.
- FIG. 7 shows a second display example of the display screen displayed by the display device 2 during endoscopic examination.
- the classification unit 34 determines that the classification score supplied from the score calculation unit 31 is greater than or equal to the first threshold value Th1, and displays the classification result indicating that the class has a lesion site. 35.
- the display control unit 35 highlights the latest image display area 70 with a bordering effect and displays a text to the effect that the lesion has been detected in order to notify the examiner of the presence of the lesion. display on the screen.
- the display control unit 35 controls the information indicating the presence of the lesion site (the latest image display area 70 in FIG. 7). border effect and text sentences). Thereby, the display control unit 35 can suitably notify the examiner that the lesion site has been detected.
- the display control unit 35 when the classification score supplied from the score calculation unit 31 is equal to or less than the second threshold Th2, the display control unit 35 notifies the examiner that there is no lesion site. As a result, the display control unit 35 can favorably prompt the examiner to shift the imaging location of the endoscope 3 (advance the operation of the endoscope 3).
- the display control unit 35 performs voice notification of the presence of a lesion site when the classification score is equal to or greater than the first threshold Th1.
- the sound output unit 16 may be instructed to output guidance or a predetermined warning sound. This also allows the display control unit 35 to allow the examiner to preferably grasp the existence of the lesion site. Similarly, the display control unit 35 instructs the sound output unit 16 to output a voice or warning sound that notifies the examiner that there is no lesion site when the classification score is equal to or less than the second threshold value Th2. good.
- FIG. 8 shows a third display example of the display screen displayed by the display device 2 during endoscopic examination.
- the display control unit 35 monitors the transition of the classification score by referring to the second calculation information storage unit D2, and determines that the classification score is on an upward trend. Accordingly, the display control unit 35 displays on the display screen the fact that the classification score is on the rise and the fact that the operation speed of the endoscope 3 is to be slowed down (that is, the imaging target is fixed as much as possible) along with the current classification score. ing.
- the display control unit 35 for example, when all of the calculated latest predetermined number of classification scores are higher than the classification score calculated at the immediately preceding processing time by a predetermined rate or more, or when the calculated latest predetermined number If the average number of classification scores is a predetermined percentage higher than their most recently calculated average of a predetermined number of classification scores, then it is determined that the classification scores are on the rise. It should be noted that the determination of an increase in the classification score is not limited to the listed examples, and may be determined by any statistical method. In this way, the display control unit 35 may output information regarding changes in classification scores. Further, the display control unit 35 may output an operation instruction for the endoscope 3 based on the information regarding the variation of the classification score.
- the display control unit 35 outputs information indicating that the classification score is on the rise when the classification score is on the rise. As a result, the display control unit 35 notifies the examiner that the classification score is increasing, and prevents detection of the lesion site due to a change in the imaging location during the determination of the presence or absence of the lesion site. can be suitably suppressed. In place of or in addition to the display control shown in the third display example, the display control unit 35 outputs a sound or a warning sound to notify the examiner that the classification score is increasing. The unit 16 may be instructed.
- FIG. 9 is an example of a flow chart executed by the image processing apparatus 1 for detection of a lesion site based on the first mode.
- the image processing apparatus 1 repeatedly executes the processing of the flowchart shown in FIG. 9 until the endoscopy is completed. Note that, for example, the image processing apparatus 1 determines that the endoscopy has ended when a predetermined input or the like to the input unit 14 or the operation unit 36 is detected.
- the captured image acquisition unit 30 of the image processing device 1 acquires the captured image Ia (step S11).
- the captured image acquisition unit 30 of the image processing apparatus 1 receives the captured image Ia from the endoscope 3 via the interface 13 .
- the display control unit 35 executes a process of displaying the photographed image Ia acquired in step S11 on the display device 2, and the like.
- the image processing device 1 calculates the likelihood ratio based on the captured image Ia acquired in step S12 and the first calculation information stored in the first calculation information storage unit D1 based on the processing in this flowchart. (step S12).
- the score calculation unit 31 of the image processing apparatus 1 acquires the captured image Ia acquired in the past in this flowchart or its feature amount or the like as the first calculation information, and obtains the first calculation information and step S12.
- the likelihood ratio is calculated based on the photographed image Ia at the current processing time. Further, the score calculation unit 31 stores the calculated likelihood ratio, the captured image Ia acquired in step S12 or its feature amount, etc. as the first calculation information in the first calculation information storage unit D1.
- the image processing device 1 calculates a classification score based on the likelihood ratio calculated in step S12 (step S13).
- the score calculation unit 31 refers to the likelihood ratios and the like stored in the first calculation information storage unit D1 to calculate the integrated likelihood ratio based on Equation (1), and the calculated integrated likelihood ratio or the integrated A function with the likelihood ratio as a variable is defined as the classification score.
- the score calculation unit 31 stores the calculated classification score and the like as second calculation information in the second calculation information storage unit D2.
- the display control unit 35 may perform a process of displaying the classification score and the like calculated by the score calculation unit 31 on the display device 2 .
- the classification unit 34 of the image processing device 1 determines whether the classification score has reached either the first threshold Th1 or the second threshold Th2 (step S14). Then, when the classification unit 34 determines that the classification score has reached one of the threshold values (Step S14; Yes), it determines (classifies) the presence or absence of the lesion site. Further, the display control unit 35 causes the display device 2 to perform display based on the determination result (classification result) of the presence or absence of the lesion site by the classification unit 34 (step S15). On the other hand, when the classification unit 34 determines that the classification score has not reached either the first threshold Th1 or the second threshold Th2 (step S14; No), the process returns to step S11. In this case, the image processing apparatus 1 further considers the captured image Ia acquired at the next processing time to update the classification score.
- FIG. 10 is an example of a flowchart executed by the image processing apparatus 1 regarding detection of a lesion site based on the second aspect.
- the image processing apparatus 1 starts the processing of the flowchart shown in FIG. 10 at predetermined time intervals until the endoscopy is completed.
- the image processing apparatus 1 starts a thread for calculating classification scores at predetermined time intervals (step S21). Then, the image processing apparatus 1 executes the same processing as steps S11 to S13 in FIG. 9 in steps S22 to S24 in each thread that has been launched. That is, the image processing apparatus 1 acquires the captured image Ia (step S22), calculates the likelihood ratio (step S23), and calculates the classification score (step S24).
- the classification unit 34 of the image processing device 1 determines whether the classification score has reached either the first threshold Th1 or the second threshold Th2 (step S25). Then, when the classification score reaches any of the thresholds (step S25; Yes), the image processing apparatus 1 performs determination (classification) regarding the presence or absence of the lesion site, and the determination result of the presence or absence of the lesion site by the classification unit 34 ( Classification result) is executed on the display device 2, and the thread started in the target flowchart is terminated (step S26). On the other hand, if the classification score has not reached either the first threshold Th1 or the second threshold Th2 (step S25; No), the process returns to step S22. In this case, in the current thread, the image processing apparatus 1 further considers the captured image Ia acquired at the next processing time to update the classification score.
- the image processing apparatus 1 may process, after the examination, a video composed of the photographed image Ia generated during the endoscopy.
- the image processing apparatus 1 performs time-series photographed images constituting the image.
- the processing of the flowchart shown in FIG. 9 or 10 is repeated for Ia until it is determined that the target video has ended.
- a target to be detected by the detection model is not limited to a lesion site, and may be any attention point that an examiner needs to pay attention to. Such points of interest are not limited to lesion sites, but include areas with inflammation, surgical scars and other cuts, folds and projections, and the tip 38 of the endoscope 3. It may also be a location on the wall surface in the lumen that is easily contacted (easily broken).
- the classification unit 34 may determine three or more types of classification instead of determining the binary classification regarding the presence or absence of the lesion site. For example, in this case, the classification unit 34 classifies “image group including lesion type 1” to “image group including lesion type X” (“X” is an integer of 2 or more) and “image group not including lesion site”. Classify X+1 types.
- the score calculation unit 31 calculates the integrated likelihood between the k-th class among the M classes and all classes other than the k-th class. Calculate the degree ratio.
- the score calculation unit 31 for example, replaces the denominators of the first and second terms on the right side of Equation (1) with the maximum likelihood of all classes other than the k-th class, and calculates the integrated likelihood ratio calculate.
- the score calculation unit 31 may calculate the integrated likelihood ratio using the sum of the likelihoods of all classes other than the k-th class instead of the maximum likelihood.
- the display control unit 35 causes the display device 2 to display information regarding the presence or absence of the lesion site and the lesion type when the lesion site is present, based on the classification result by the classification unit 34 .
- the image processing apparatus 1 can favorably assist diagnosis by the examiner.
- FIG. 11 is a block diagram of an image processing device 1X according to the second embodiment.
- the image processing device 1X includes acquisition means 30X, score calculation means 31X, and classification means 34X.
- the image processing device 1X may be composed of a plurality of devices.
- the acquisition means 30X acquires a photographed image of the inspection object photographed by the photographing unit provided in the endoscope.
- the acquisition means 30X may immediately acquire the captured image generated by the imaging unit, or may acquire, at a predetermined timing, the captured image generated by the imaging unit and stored in the storage device in advance.
- the acquisition unit 30X can be, for example, the captured image acquisition unit 30 in the first embodiment (including modifications; the same applies hereinafter).
- the score calculation means 31X calculates a score regarding the likelihood that a spot of interest exists in the captured image based on a plurality of time-series captured images.
- the score calculator 31X can be the score calculator 31 in the first embodiment, for example.
- the classification means 34X classifies a plurality of captured images based on the scores.
- the classifying means 34X can be, for example, the classifying section 34 in the first embodiment.
- FIG. 12 is an example of a flowchart showing a processing procedure in the second embodiment.
- the acquiring means 30X acquires a photographed image of an inspection object photographed by the photographing unit provided in the endoscope (step S31).
- the score calculation unit 31X calculates a score regarding the likelihood that a spot of interest exists in a captured image based on a plurality of captured images in time series (step S32).
- the classification means 34X classifies the plurality of captured images based on the scores (step S33).
- the image processing device 1X can accurately detect a point of interest by calculating a score regarding the likelihood that a point of interest is present based on a plurality of time-series captured images. can.
- Non-transitory computer readable media include various types of tangible storage media.
- Examples of non-transitory computer-readable media include magnetic storage media (e.g., floppy disks, magnetic tapes, hard disk drives), magneto-optical storage media (e.g., magneto-optical discs), CD-ROMs (Read Only Memory), CD-Rs, CD-R/W, semiconductor memory (eg mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM, RAM (Random Access Memory)).
- the program may also be delivered to the computer on various types of transitory computer readable medium.
- Examples of transitory computer-readable media include electrical signals, optical signals, and electromagnetic waves.
- Transitory computer-readable media can deliver the program to the computer via wired channels, such as wires and optical fibers, or wireless channels.
- the score calculation means updates the score based on the captured image acquired by the acquisition means and the captured image used for calculating the score until the score reaches a predetermined threshold; the classifier determines the classification if the score reaches the threshold; The image processing apparatus according to claim 1 or 2.
- [Appendix 4] 4. The image processing according to claim 3, wherein, when the score reaches the threshold, the score calculation means calculates the score based on the photographed image acquired by the acquisition means after the score reaches the threshold.
- Device [Appendix 5] The score calculation means sets a start point of the photographed images used for calculating the score at predetermined time intervals, and based on the photographed images acquired by the acquisition means after the start point for each of the set start points, 4.
- the image processing device according to claim 1, wherein said score is calculated.
- Appendix 6 6.
- the image processing apparatus according to any one of claims 1 to 5, further comprising output control means for outputting information on said score and classification result by said classification means by means of a display device or a sound output device.
- Appendix 7 7.
- the output control unit according to claim 6, wherein, when the plurality of captured images are classified according to the existence of the attention point, the output control means outputs information indicating that the attention point exists from the display device or the sound output device.
- the described image processing device [Appendix 8] 8.
- the output control means outputs information indicating that the classification is undecided from the display device or the sound output device.
- Appendix 11 Acquiring a photographed image of an object to be inspected by an imaging unit provided in the endoscope, Based on the plurality of time-series captured images, calculating a score regarding the likelihood that a spot of interest exists in the captured image,
- a storage medium storing a program for causing a computer to execute processing for classifying the plurality of captured images based on the scores.
- Reference Signs List 1 1X image processing device 2 display device 3 endoscope 4 server device 11 processor 12 memory 13 interface 14 input unit 15 light source unit 16 sound output unit 100 endoscopy system
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Surgery (AREA)
- Pathology (AREA)
- Optics & Photonics (AREA)
- Biophysics (AREA)
- Signal Processing (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Animal Behavior & Ethology (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Endoscopes (AREA)
Abstract
Description
内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得する取得手段と、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出するスコア算出手段と、
前記スコアに基づき、前記複数の撮影画像を分類する分類手段と、
を有する画像処理装置である。
コンピュータが、
内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得し、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出し、
前記スコアに基づき、前記複数の撮影画像を分類する、
画像処理方法である。
内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得し、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出し、
前記スコアに基づき、前記複数の撮影画像を分類する処理をコンピュータに実行させるプログラムを格納した記憶媒体である。
(1)システム構成
図1は、内視鏡検査システム100の概略構成を示す。図1に示すように、内視鏡検査システム100は、内視鏡を利用した検査又は治療を行う医師等の検査者に対して病変の疑いがある部位(病変部位)を提示するシステムであって、主に、画像処理装置1と、表示装置2と、画像処理装置1に接続された内視鏡スコープ3と、を備える。
(a)頭頚部:咽頭ガン、悪性リンパ腫、乳頭腫
(b)食道:食道ガン、食道炎、食道裂孔ヘルニア、バレット食道、食道静脈瘤、食道アカラシア、食道粘膜下腫瘍、食道良性腫瘍
(c)胃:胃ガン、胃炎、胃潰瘍、胃ポリープ、胃腫瘍
(d)十二指腸:十二指腸ガン、十二指腸潰瘍、十二指腸炎、十二指腸腫瘍、十二指腸リンパ腫
(e)小腸:小腸ガン、小腸腫瘍性疾患、小腸炎症性疾患、小腸血管性疾患
(f)大腸:大腸ガン、大腸腫瘍性疾患、大腸炎症性疾患、大腸ポリープ、大腸ポリポーシス、クローン病、大腸炎、腸結核、痔
図2は、画像処理装置1のハードウェア構成を示す。画像処理装置1は、主に、プロセッサ11と、メモリ12と、インターフェース13と、入力部14と、光源部15と、音出力部16と、を含む。これらの各要素は、データバス19を介して接続されている。
次に、画像処理装置1による病変部位の検出処理について説明する。概略的には、画像処理装置1は、可変枚数の時系列の撮影画像Iaに基づき病変部位の検出を行う。これにより、画像処理装置1は、単一画像からの判別が難しい病変部位についても的確に病変部位を検出し、その検出結果の提示を行う。なお、単一画像からの判別が難しい病変部位は、例えば、バレット食道等における、非癌化部分と平坦な癌化部分などが存在する。
次に、表示制御部35が実行する表示装置2の表示制御について説明する。
図9は、第1態様に基づく病変部位の検出に関して画像処理装置1が実行するフローチャートの一例である。画像処理装置1は、図9に示すフローチャートの処理を、内視鏡検査の終了まで繰り返し実行する。なお、例えば、画像処理装置1は、入力部14又は操作部36への所定の入力等を検知した場合に、内視鏡検査が終了したと判定する。
次に、上述した実施形態に好適な変形例について説明する。以下の変形例は、組み合わせて上述の実施形態に適用してもよい。
画像処理装置1は、内視鏡検査時に生成された撮影画像Iaから構成された映像を、検査後において処理してもよい。
検知モデルにより検知する対象は、病変部位に限らず、検査者が注目する必要がある任意の注目箇所であってもよい。このような注目箇所は、病変部位に限らず、炎症が生じている箇所、手術痕その他の切り傷が生じている箇所、ひだや突起が生じている箇所、内視鏡スコープ3の先端部38が管腔内の壁面において接触しやすい(閊えやすい)箇所などであってもよい。
分類部34は、病変部位の存否に関する2値の分類を決定する代わりに、3種類以上の分類を決定してもよい。例えば、この場合、分類部34は、「病変タイプ1を含む画像群」~「病変タイプXを含む画像群」(「X」は2以上の整数)及び「病変部位を含まない画像群」のX+1種類を分類する。
図11は、第2実施形態における画像処理装置1Xのブロック図である。画像処理装置1Xは、取得手段30Xと、スコア算出手段31Xと、分類手段34Xとを備える。画像処理装置1Xは、複数の装置から構成されてもよい。
内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得する取得手段と、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出するスコア算出手段と、
前記スコアに基づき、前記複数の撮影画像を分類する分類手段と、
を有する画像処理装置。
[付記2]
前記スコア算出手段は、前記複数の撮影画像における前記注目箇所の存否に関する尤度比に基づき、前記スコアを算出し、
前記分類手段は、前記複数の撮影画像における前記注目箇所の存否に関する分類を行う、請求項1に記載の画像処理装置。
[付記3]
前記スコア算出手段は、前記スコアが所定の閾値に達するまでの間、前記取得手段が取得する撮影画像と前記スコアの算出に用いた撮影画像とに基づき、前記スコアを更新し、
前記分類手段は、前記スコアが前記閾値に達した場合に、前記分類を決定する、
請求項1または2に記載の画像処理装置。
[付記4]
前記スコア算出手段は、前記スコアが前記閾値に達した場合、前記スコアが前記閾値に達した後に前記取得手段が取得する撮影画像に基づき、前記スコアを算出する、請求項3に記載の画像処理装置。
[付記5]
前記スコア算出手段は、所定時間間隔により、前記スコアの算出に用いる撮影画像の開始時点を設定し、設定した前記開始時点ごとに、当該開始時点以後に前記取得手段が取得する撮影画像に基づき、前記スコアを算出する、請求項1~3のいずれか一項に記載の画像処理装置。
[付記6]
前記スコア及び前記分類手段による分類結果に関する情報を表示装置又は音出力装置により出力する出力制御手段を有する、請求項1~5のいずれか一項に記載の画像処理装置。
[付記7]
前記出力制御手段は、前記注目箇所が存在すると前記複数の撮影画像が分類された場合、前記注目箇所が存在することを示す情報を前記表示装置又は前記音出力装置により出力する、請求項6に記載の画像処理装置。
[付記8]
前記出力制御手段は、前記分類手段による分類が未決定の場合、前記分類が未決定であることを示す情報を前記表示装置又は前記音出力装置により出力する、請求項6または7に記載の画像処理装置。
[付記9]
前記出力制御手段は、前記スコアが上昇傾向にある場合、前記スコアが上昇傾向にあることを示す情報を前記表示装置又は前記音出力装置により出力する、請求項6~8のいずれか一項に記載の画像処理装置。
[付記10]
コンピュータが、
内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得し、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出し、
前記スコアに基づき、前記複数の撮影画像を分類する、
画像処理方法。
[付記11]
内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得し、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出し、
前記スコアに基づき、前記複数の撮影画像を分類する処理をコンピュータに実行させるプログラムを格納した記憶媒体。
2 表示装置
3 内視鏡スコープ
4 サーバ装置
11 プロセッサ
12 メモリ
13 インターフェース
14 入力部
15 光源部
16 音出力部
100 内視鏡検査システム
Claims (11)
- 内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得する取得手段と、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出するスコア算出手段と、
前記スコアに基づき、前記複数の撮影画像を分類する分類手段と、
を有する画像処理装置。 - 前記スコア算出手段は、前記複数の撮影画像における前記注目箇所の存否に関する尤度比に基づき、前記スコアを算出し、
前記分類手段は、前記複数の撮影画像における前記注目箇所の存否に関する分類を行う、請求項1に記載の画像処理装置。 - 前記スコア算出手段は、前記スコアが所定の閾値に達するまでの間、前記取得手段が取得する撮影画像と前記スコアの算出に用いた撮影画像とに基づき、前記スコアを更新し、
前記分類手段は、前記スコアが前記閾値に達した場合に、前記分類を決定する、
請求項1または2に記載の画像処理装置。 - 前記スコア算出手段は、前記スコアが前記閾値に達した場合、前記スコアが前記閾値に達した後に前記取得手段が取得する撮影画像に基づき、前記スコアを算出する、請求項3に記載の画像処理装置。
- 前記スコア算出手段は、所定時間間隔により、前記スコアの算出に用いる撮影画像の開始時点を設定し、設定した前記開始時点ごとに、当該開始時点以後に前記取得手段が取得する撮影画像に基づき、前記スコアを算出する、請求項1~3のいずれか一項に記載の画像処理装置。
- 前記スコア及び前記分類手段による分類結果に関する情報を表示装置又は音出力装置により出力する出力制御手段を有する、請求項1~5のいずれか一項に記載の画像処理装置。
- 前記出力制御手段は、前記注目箇所が存在すると前記複数の撮影画像が分類された場合、前記注目箇所が存在することを示す情報を前記表示装置又は前記音出力装置により出力する、請求項6に記載の画像処理装置。
- 前記出力制御手段は、前記分類手段による分類が未決定の場合、前記分類が未決定であることを示す情報を前記表示装置又は前記音出力装置により出力する、請求項6または7に記載の画像処理装置。
- 前記出力制御手段は、前記スコアが上昇傾向にある場合、前記スコアが上昇傾向にあることを示す情報を前記表示装置又は前記音出力装置により出力する、請求項6~8のいずれか一項に記載の画像処理装置。
- コンピュータが、
内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得し、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出し、
前記スコアに基づき、前記複数の撮影画像を分類する、
画像処理方法。 - 内視鏡に設けられた撮影部により検査対象を撮影した撮影画像を取得し、
時系列となる複数の前記撮影画像に基づき、前記撮影画像に注目すべき注目箇所が存在する尤もらしさに関するスコアを算出し、
前記スコアに基づき、前記複数の撮影画像を分類する処理をコンピュータに実行させるプログラムを格納した記憶媒体。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP21937937.7A EP4327721A4 (en) | 2021-04-23 | 2021-04-23 | IMAGE PROCESSING DEVICE, IMAGE PROCESSING METHOD AND STORAGE MEDIUM |
US18/287,543 US20240203086A1 (en) | 2021-04-23 | 2021-04-23 | Image processing device, image processing method, and storage medium |
PCT/JP2021/016473 WO2022224446A1 (ja) | 2021-04-23 | 2021-04-23 | 画像処理装置、画像処理方法及び記憶媒体 |
JP2023516007A JPWO2022224446A5 (ja) | 2021-04-23 | 画像処理装置、画像処理方法及び、プログラム |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2021/016473 WO2022224446A1 (ja) | 2021-04-23 | 2021-04-23 | 画像処理装置、画像処理方法及び記憶媒体 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022224446A1 true WO2022224446A1 (ja) | 2022-10-27 |
Family
ID=83723267
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/016473 WO2022224446A1 (ja) | 2021-04-23 | 2021-04-23 | 画像処理装置、画像処理方法及び記憶媒体 |
Country Status (3)
Country | Link |
---|---|
US (1) | US20240203086A1 (ja) |
EP (1) | EP4327721A4 (ja) |
WO (1) | WO2022224446A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024075242A1 (ja) * | 2022-10-06 | 2024-04-11 | 日本電気株式会社 | 画像処理装置、画像処理方法及び記憶媒体 |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016110993A1 (ja) * | 2015-01-09 | 2016-07-14 | オリンパス株式会社 | 内視鏡システム、内視鏡装置及び内視鏡システムの制御方法 |
WO2019138772A1 (ja) * | 2018-01-10 | 2019-07-18 | 富士フイルム株式会社 | 画像処理装置、プロセッサ装置、画像処理方法、及びプログラム |
JP2019180966A (ja) * | 2018-04-13 | 2019-10-24 | 学校法人昭和大学 | 内視鏡観察支援装置、内視鏡観察支援方法、及びプログラム |
WO2020003607A1 (ja) | 2018-06-25 | 2020-01-02 | オリンパス株式会社 | 情報処理装置、モデル学習方法、データ認識方法および学習済みモデル |
WO2020165978A1 (ja) * | 2019-02-13 | 2020-08-20 | オリンパス株式会社 | 画像記録装置、画像記録方法および画像記録プログラム |
WO2020170791A1 (ja) * | 2019-02-19 | 2020-08-27 | 富士フイルム株式会社 | 医療画像処理装置及び方法 |
WO2020194497A1 (ja) | 2019-03-26 | 2020-10-01 | 日本電気株式会社 | 情報処理装置、個人識別装置、情報処理方法及び記憶媒体 |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110301447A1 (en) * | 2010-06-07 | 2011-12-08 | Sti Medical Systems, Llc | Versatile video interpretation, visualization, and management system |
-
2021
- 2021-04-23 WO PCT/JP2021/016473 patent/WO2022224446A1/ja active Application Filing
- 2021-04-23 EP EP21937937.7A patent/EP4327721A4/en active Pending
- 2021-04-23 US US18/287,543 patent/US20240203086A1/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016110993A1 (ja) * | 2015-01-09 | 2016-07-14 | オリンパス株式会社 | 内視鏡システム、内視鏡装置及び内視鏡システムの制御方法 |
WO2019138772A1 (ja) * | 2018-01-10 | 2019-07-18 | 富士フイルム株式会社 | 画像処理装置、プロセッサ装置、画像処理方法、及びプログラム |
JP2019180966A (ja) * | 2018-04-13 | 2019-10-24 | 学校法人昭和大学 | 内視鏡観察支援装置、内視鏡観察支援方法、及びプログラム |
WO2020003607A1 (ja) | 2018-06-25 | 2020-01-02 | オリンパス株式会社 | 情報処理装置、モデル学習方法、データ認識方法および学習済みモデル |
WO2020165978A1 (ja) * | 2019-02-13 | 2020-08-20 | オリンパス株式会社 | 画像記録装置、画像記録方法および画像記録プログラム |
WO2020170791A1 (ja) * | 2019-02-19 | 2020-08-27 | 富士フイルム株式会社 | 医療画像処理装置及び方法 |
WO2020194497A1 (ja) | 2019-03-26 | 2020-10-01 | 日本電気株式会社 | 情報処理装置、個人識別装置、情報処理方法及び記憶媒体 |
Non-Patent Citations (1)
Title |
---|
See also references of EP4327721A4 |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024075242A1 (ja) * | 2022-10-06 | 2024-04-11 | 日本電気株式会社 | 画像処理装置、画像処理方法及び記憶媒体 |
WO2024075410A1 (ja) * | 2022-10-06 | 2024-04-11 | 日本電気株式会社 | 画像処理装置、画像処理方法及び記憶媒体 |
Also Published As
Publication number | Publication date |
---|---|
EP4327721A4 (en) | 2024-06-19 |
EP4327721A1 (en) | 2024-02-28 |
US20240203086A1 (en) | 2024-06-20 |
JPWO2022224446A1 (ja) | 2022-10-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110049709B (zh) | 图像处理装置 | |
JP5276225B2 (ja) | 医用画像処理装置及び医用画像処理装置の作動方法 | |
EP3777645A1 (en) | Endoscope observation assistance device, endoscope observation assistance method, and program | |
JP7413585B2 (ja) | 医用画像処理装置、作動方法、プログラム、及び記録媒体、診断支援装置並びに内視鏡システム | |
JP2011234931A (ja) | 画像処理装置、画像処理方法、および画像処理プログラム | |
JP2008029520A (ja) | 医療用画像処理装置及び医療用画像処理方法 | |
JP4749732B2 (ja) | 医用画像処理装置 | |
WO2020054543A1 (ja) | 医療画像処理装置及び方法、内視鏡システム、プロセッサ装置、診断支援装置並びにプログラム | |
WO2022224446A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
US20190298159A1 (en) | Image processing device, operation method, and computer readable recording medium | |
JP2019111040A (ja) | 画像解析装置及び画像解析方法 | |
WO2023042273A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2023162216A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2022180786A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
JP2020175051A (ja) | 診断支援システム、及び診断支援用プログラム | |
US20240016366A1 (en) | Image diagnosis system for lesion | |
WO2024075240A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2023187886A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2024013848A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2022185369A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2023233453A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2023181353A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
WO2024075242A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 | |
JP7264407B2 (ja) | 訓練用の大腸内視鏡観察支援装置、作動方法、及びプログラム | |
WO2024084578A1 (ja) | 画像処理装置、画像処理方法及び記憶媒体 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21937937 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2023516007 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18287543 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2021937937 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2021937937 Country of ref document: EP Effective date: 20231123 |