CN108648195B - Image processing method and device - Google Patents
Image processing method and device Download PDFInfo
- Publication number
- CN108648195B CN108648195B CN201810437266.0A CN201810437266A CN108648195B CN 108648195 B CN108648195 B CN 108648195B CN 201810437266 A CN201810437266 A CN 201810437266A CN 108648195 B CN108648195 B CN 108648195B
- Authority
- CN
- China
- Prior art keywords
- sequence
- slice
- features
- image
- feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses an image processing method and device, wherein the method comprises the following steps: acquiring a plurality of continuous slice images corresponding to a target object to form a slice image sequence; extracting sequence features of the sequence of slice images and extracting internal features of the slice images, wherein the sequence features are features characterizing association among a plurality of the slice images; and acquiring a segmentation map of the target object according to the sequence feature and the internal feature. The method can be used for carrying out image segmentation on a plurality of objects gathered together, can be used for carrying out more complete image display on the characteristics of each target object, is higher in quality of the formed segmentation graph, and improves the efficiency of analyzing the target objects by utilizing the segmentation graph of the target objects.
Description
Technical Field
The present invention relates to the field of image processing, and in particular, to an image processing method and apparatus.
Background
In the field of image processing, especially when multiple objects are gathered together, it is very difficult to extract a segmentation map with high quality of one or more target objects, which results in poor quality of the obtained segmentation map, and the characteristics of the target objects cannot be fully reflected, thereby affecting the analysis of the target objects based on the segmentation map of the target objects. For example, when a medical instrument is used to perform image analysis on each organ inside the human body, it is very difficult to obtain a segmentation map having a high quality for each organ because a plurality of organs of the human body are close to each other or are superimposed one on another.
Disclosure of Invention
An object of embodiments of the present invention is to provide an image processing method and apparatus, which can perform image segmentation on a plurality of objects that are grouped together, and can perform more complete image display on features of each target object, thereby improving efficiency of analyzing the target object by using a segmentation map of the target object.
In order to solve the technical problem, the embodiment of the invention adopts the following technical scheme: an image processing method, comprising:
acquiring a plurality of continuous slice images corresponding to a target object to form a slice image sequence;
extracting sequence features of the slice image sequence and extracting internal features of the slice images, wherein the sequence features are features representing association among a plurality of the slice images;
and acquiring a segmentation map of the target object according to the sequence feature and the internal feature.
Preferably, the method further comprises:
performing an extraction operation of extracting features on the slice image sequence to generate a first feature sequence;
performing at least one compression operation on the first feature sequence to reduce the operation amount when the sequence features and the internal features are extracted;
And after each compression operation, the extraction operation of the sequence features and the internal features is carried out on the first feature sequence.
Preferably, the slice image sequence includes a center slice, and the extracting the sequence feature of the slice image sequence includes:
and respectively carrying out extraction operation of the sequence characteristics on the forward subsequence and the backward subsequence of the central slice, and correlating the two operation results to form a forward and backward subsequence set.
Preferably, the method further comprises:
performing at least one decompression operation on the slice image sequence after the extraction operation is performed to restore the slice image sequence;
and after each decompression operation, performing the sequence feature extraction operation on the forward and backward subsequence sets, and performing the internal feature extraction operation on the central slice.
Preferably, the slice image sequence includes a center slice, and the acquiring a plurality of consecutive slice images corresponding to the target object to form the slice image sequence includes:
determining the central slice, selecting a forward subsequence and a backward subsequence of the central slice according to a preset specification according to the shape of a target object, and keeping the sample balance of the slice image sequence;
Preferably, the method further comprises:
and performing image preprocessing on the slice image according to the physical characteristics of the target object so as to avoid the interference of images of other objects, wherein the image preprocessing comprises the step of adaptively changing the image attribute of the tangent plane image.
The embodiment of the application further provides an image processing device, which comprises an acquisition module, a calculation module and a generation module:
the acquisition module is configured to acquire a plurality of continuous slice images corresponding to a target object to form a slice image sequence;
the computing module is configured to extract sequence features of the sequence of slice images and extract internal features of the slice images, wherein the sequence features are features characterizing associations between a plurality of the slice images;
the generation module is configured to obtain a segmentation map of the target object from the sequence features and the internal features.
Preferably, the calculation module comprises an extraction unit and a compression unit;
the extraction unit is configured to perform extraction operation of extracting features on the slice image sequence to generate a first feature sequence;
the compression unit is configured to perform at least one compression operation on the first feature sequence to reduce an operation amount when the extraction unit extracts the sequence feature and the internal feature;
The extraction unit is further configured to perform the extraction operation of the sequence feature and the internal feature on the first feature sequence after the compression unit performs each compression operation.
Preferably, the slice image sequence includes a center slice, and the computing module is further configured to perform the sequence feature extraction operation on a forward subsequence and a backward subsequence of the center slice, respectively, and correlate the two operation results to form a forward and backward subsequence set.
Preferably, the calculation module comprises an extraction unit and a decompression unit;
the decompression unit is configured to perform at least one decompression operation on the slice image sequence after the extraction operation is performed so as to restore the slice image sequence;
the extraction unit is configured to perform the sequence feature extraction operation on the forward and backward subsequence sets and perform the internal feature extraction operation on the central slice after the decompression unit performs each decompression operation.
The embodiment of the invention has the beneficial effects that: the method can be used for carrying out image segmentation on a plurality of objects gathered together, can be used for carrying out more complete image display on the characteristics of each target object, is higher in quality of the formed segmentation graph, and improves the efficiency of analyzing the target objects by utilizing the segmentation graph of the target objects.
Drawings
FIG. 1 is a flow chart of an image processing method according to an embodiment of the present invention;
FIG. 2 is a flowchart of an embodiment of an image processing method according to the invention;
FIG. 3 is a flowchart of another embodiment of an image processing method according to the present invention;
fig. 4 is a schematic structural diagram of an image processing apparatus according to an embodiment of the present invention.
Detailed Description
Various aspects and features of the present invention are described herein with reference to the drawings.
It will be understood that various modifications may be made to the embodiments of the invention herein. Accordingly, the foregoing description should not be construed as limiting, but merely as exemplifications of embodiments. Other modifications will occur to those skilled in the art which are within the scope and spirit of the invention.
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and, together with a general description of the invention given above, and the detailed description of the embodiments given below, serve to explain the principles of the invention.
These and other characteristics of the invention will become apparent from the following description of a preferred form of embodiment, given as a non-limiting example, with reference to the accompanying drawings.
It should also be understood that, although the invention has been described with reference to some specific examples, a person of skill in the art shall certainly be able to achieve many other equivalent forms of the invention, having the characteristics as set forth in the claims and hence all coming within the field of protection defined thereby.
The above and other aspects, features and advantages of the present invention will become more apparent in view of the following detailed description when taken in conjunction with the accompanying drawings.
Specific embodiments of the present invention are described hereinafter with reference to the accompanying drawings; however, it is to be understood that the disclosed embodiments are merely exemplary of the invention, which can be embodied in various forms. Well-known and/or repeated functions and constructions are not described in detail to avoid obscuring the invention in unnecessary or unnecessary detail. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a basis for the claims and as a representative basis for teaching one skilled in the art to variously employ the present invention in virtually any appropriately detailed structure.
The specification may use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which may each refer to one or more of the same or different embodiments in accordance with the invention.
The image processing method provided by the embodiment of the invention can be used for carrying out image segmentation on a plurality of objects which are gathered together to obtain the relatively independent segmentation maps of the target object, so that a user can carry out more detailed operations such as analysis or inspection on the target object according to the segmentation maps, and the user can conveniently process the segmentation maps by using other methods on the basis of the segmentation maps of the target object. As shown in fig. 1, the method comprises the steps of:
s1, a plurality of consecutive slice images corresponding to the target object are acquired to form a slice image sequence. The data sampling can be carried out on the area where the target object is located, a plurality of continuous slice images (frames) are extracted, the slice images are mutually parallel, the regularity of the slice images is guaranteed, and in addition, the number of the slice images can be specifically set according to the characteristics of the processing object. The plurality of consecutive slice images form a slice image sequence, which may represent a set of a plurality of consecutive slice images, which may be utilized in processing the target object, and one or more slice image sequences may be associated with one target object, such as extracting consecutive slice images from different angles from the target object, and in one embodiment, the slice image sequence may include a center slice, which is a main slice image of the target object and carries core data of the target object.
And S2, extracting sequence characteristics of the slice image sequence and extracting internal characteristics of the slice images, wherein the sequence characteristics are characteristics for representing the correlation among a plurality of slice images. The related data between the slice images can be obtained from the sequence characteristics, and the related data of the target object in a plurality of slice images can be effectively utilized, so that all the related data of the target object can be effectively utilized, and the image processing is more accurate. In one embodiment, the extraction of the sequence features may be performed one or more times on the sequence of slice images. The internal features of the slice images can represent the data features of the corresponding slice images, and in one embodiment, the internal features of each slice image can be extracted; in another embodiment, the extraction of the internal features may be performed once or more times for each slice image; in yet another embodiment, the extraction of the internal features of the slice image may be performed using a convolution operation.
And S3, acquiring a segmentation map of the target object according to the sequence features and the internal features. Specifically, the corresponding sequence features and internal features may be connected and fused according to the type or specific features of the target object, for example, different features may be combined to obtain a segmentation map of the target object, where the segmentation map is an image (or data) of the target object, and the segmentation map enables the target object to be separated from other multiple objects, so that the target object may be further operated according to the segmentation map, for example, the segmentation map may be directly utilized to determine aspects such as a form and a state of the target object, or reprocessed based on the segmentation map, so that a processing result is also relatively accurate. It should be noted that the method can be applied to a medical apparatus for processing an image of a human body, for example, processing a plurality of organs (usually gathered together) in the human body, and acquiring a first segmentation map of a first organ, so that the first organ can be relatively independently displayed; or selecting a predetermined number of different organs (such as the second organ and the third organ) from the plurality of organs, and acquiring a corresponding second segmentation map so that the predetermined number of different organs are relatively independently displayed.
In one embodiment of the present application, as shown in fig. 2, the image processing method further includes the steps of:
and S4, performing extraction operation of the extracted features on the slice image sequence to generate a first feature sequence. After a plurality of continuous slice images are acquired, the formed slice image sequence is subjected to extraction operation, and the generated first feature sequence may include the same number of first slice data units as the slice images, and the first slice data units may also be displayed in the form of images or appear in the form of data units. Also in this embodiment the extraction operation may be performed using a separate sequence feature extractor.
And S5, performing at least one compression operation on the first feature sequence to reduce the operation amount when the sequence features and the internal features are extracted. Specifically, the compression operation (encoding) can effectively reduce the original data volume under the condition of ensuring data security. For example, the original 512 × 64 pixel image in the first feature sequence is converted into 256 × 128 pixel image or data unit after compression operation, and then at least one operation of extracting the sequence feature and the internal feature is performed, so that the operation cup for extracting the sequence feature and the internal feature is effectively reduced due to the reduction of the data amount. In addition, a second compression operation (encoding) can be performed, for example, after the image or data unit with 256 × 128 pixels is subjected to the compression operation, the image or data unit with 256 × 128 pixels is converted into the image or data unit with 128 × 256 pixels, and then at least one operation of extracting the sequence feature and the internal feature is performed. In addition, in one embodiment, the compression operation (encoding) may operate using a Max-pooling layer (Max-pooling layer).
And S6, after each compression operation, carrying out extraction operation of the sequence features and the internal features on the first feature sequence. The first feature sequence may be subjected to a predetermined number of compression operations for a plurality of times according to actual conditions, and after each compression operation, a predetermined number of operations for extracting sequence features and internal features may be performed. Therefore, the extraction operation of the sequence features and the internal features is carried out for a plurality of times in the whole process, and the accuracy of the data of the target object in the data processing process is ensured.
In one embodiment of the present application, the slice image sequence includes a center slice, and the step of extracting the sequence feature of the slice image sequence includes: and respectively carrying out extraction operation of sequence characteristics on the forward subsequence and the backward subsequence of the central slice, and correlating the two operation results to form a forward and backward subsequence set. The central slice is a main slice image of the target object and carries core data of the target object. The central slice corresponds to the forward subsequence and the backward subsequence, and the forward subsequence and the backward subsequence can be regarded as wrapping the central slice therein from two sides respectively. In one embodiment, the priority of the slice image closer to the central slice is higher, and the priority of the slice image farther from the central slice is lower, so that the extraction operation of the sequence features of the forward subsequence can be performed based on the priority, and the extraction operation of the sequence features of the backward subsequence is performed, so that the slice image closer to the core data plays a larger role, and the accuracy in data processing is improved. Further, performing the above steps may be performed using a separate sequence feature extractor. After the extraction operation is performed, the two operation results can be associated to form a forward and backward subsequence set, and then the forward and backward subsequence set can be associated with the center slice again so as to perform unified processing.
In one embodiment of the present application, as shown in fig. 3, the image processing method further includes the steps of:
s7, the slice image sequence after the extraction operation is decompressed at least once to restore the slice image sequence. Specifically, after the original slice image sequence is subjected to one or more extraction operations (sequence features and internal features are extracted, and multiple compression operations are also performed), the slice image sequence may be restored by performing decompression operations for corresponding times. For example, a decompression operation is performed, and the original 128 × 256 pixel image or data unit is converted into 256 × 128 pixel image or data unit after the decompression operation; and performing decompression operation again, converting the 256 × 128 pixel image into 512 × 64 pixel image or data unit after the decompression operation, and finally restoring the slice image sequence.
And S8, after each decompression operation, performing sequence feature extraction operation on the forward and backward subsequence sets, and performing internal feature extraction operation on the central slice. Specifically, after the operation is carried out, the data is not abnormal in the reduction process, and the accuracy of the data is ensured. Furthermore, after each decompression operation, the operated forward and backward subsequence sets can be associated with the center slice so as to perform the next sequence feature extraction operation and internal feature extraction operation. In addition, the decompression operation may be performed by using a deconvolution operation.
In an embodiment of the present application, the slice image sequence includes a center slice, and the step of acquiring a plurality of consecutive slice images corresponding to the target object to form the slice image sequence includes: determining a central slice, selecting a forward subsequence and a backward subsequence of the central slice according to the form of the target object and a preset rule, and keeping the sample balance of the slice image sequence. Specifically, the central slice may be determined according to the general shape of the target object or a position point of interest of a user on the target object, and the forward subsequence and the backward subsequence of the central slice may be selected according to a preset rule, for example, all slice images in which the target object appears may be extracted during sampling, and then a part of slice images in which the target object does not appear may be randomly extracted from the slice images in which the target object does not appear, with the central slice as a reference, so as to ensure sample balance of the slice image sequence.
In one embodiment of the present application, the image processing method further includes: and performing image preprocessing on the slice image according to the physical characteristics of the target object to avoid the interference of the images of other objects, wherein the image preprocessing comprises the step of adaptively changing the image attribute of the section image. For example, if image segmentation is required for each organ in the abdomen when the human body image is processed, the original slice image may be subjected to adaptive modification of image attributes, such as adjusting an image with a gray value less than-75 in the original slice image to-75 and an image with a gray value greater than 175 in the original slice image to 175, so that the image segmentation can be more adaptive to each organ in the abdomen of the human body, and the generated segmentation map of one or more organs has higher quality.
The embodiment of the present application further provides an image processing apparatus, which can perform image segmentation on a plurality of objects aggregated together to obtain a relatively independent segmentation map of a target object, so that a user can perform operations such as more detailed analysis or inspection on the target object according to the segmentation map, and the user can further process the segmentation map on the basis of the segmentation map of the target object. As shown in fig. 4, the image processing apparatus includes an acquisition module, a calculation module, and a generation module.
The acquisition module is configured to acquire a plurality of consecutive slice images corresponding to the target object to form a sequence of slice images. The acquisition module can carry out data sampling on the region where the target object is located, and extract a plurality of continuous slice images (frames), wherein the slice images are mutually parallel, so that the regularity of the slice images is ensured, and in addition, the number of the slice images can also be specifically set according to the characteristics of the processing object. The plurality of consecutive slice images form a slice image sequence, which may represent a set of a plurality of consecutive slice images, which may be utilized in processing the target object, and furthermore, one target object may correspond to one or more slice image sequences, such as for extracting consecutive slice images from different angles to the target object, and in one embodiment, the slice image sequence may include a center slice, which is a main slice image of the target object and carries core data of the target object.
The computing module is configured to extract sequence features of the sequence of slice images and extract internal features of the slice images, wherein the sequence features are features characterizing an association between the plurality of slice images. The related data between the slice images can be obtained from the sequence characteristics, and the related data of the target object in a plurality of slice images can be effectively utilized, so that all the related data of the target object can be effectively utilized, and the image processing is more accurate. In one embodiment, the computation module may perform one or more extractions of sequence features on the sequence of slice images. The internal features of the slice images can represent the data features of the corresponding slice images, and in one embodiment, the computing module can extract the internal features of each slice image; in another embodiment, the computing module may perform the extraction of the internal features for each slice image one or more times; in yet another embodiment, the computation module may utilize convolution operations to perform the extraction of the internal features of the slice images.
The generation module is configured to obtain a segmentation map of the target object from the sequence features and the internal features. Specifically, the generating module may perform connection fusion on the corresponding sequence features and internal features according to the type or specific features of the target object, for example, combine different features to obtain a segmentation map of the target object, where the segmentation map is an image (or data) of the target object, and the segmentation map enables the target object to be separated from other multiple objects, so that the target object can be further operated according to the segmentation map, for example, the segmentation map is directly utilized to determine aspects such as a form and a state of the target object, or reprocessed based on the segmentation map, so that a processing result is also relatively accurate. It should be noted that the apparatus can be applied to a medical device for processing an image of a human body, for example, processing a plurality of organs (usually gathered together) in the human body, and acquiring a first segmentation map of a first organ, so that the first organ can be relatively independently displayed; or selecting a predetermined number of different organs (such as the second organ and the third organ) from the plurality of organs, and acquiring a corresponding second segmentation map so that the predetermined number of different organs are relatively independently displayed.
In one embodiment of the present application, in conjunction with FIG. 4, the computation module includes an extraction unit and a compression unit.
The extraction unit is configured to perform an extraction operation of extracting features on the slice image sequence to generate a first feature sequence. After acquiring a plurality of continuous slice images, performing an extraction operation on the formed slice image sequence, and generating a first feature sequence which may include the same number of first slice data units as the slice images, where the first slice data units may also be displayed in the form of images or may appear in the form of data units. Also in this embodiment the extraction operation may be performed using a separate sequence feature extractor.
The compression unit is configured to perform at least one compression operation on the first feature sequence to reduce an operation amount when the extraction unit extracts the sequence feature and the internal feature. Specifically, the compression operation (encoding) of the compression unit can effectively reduce the original data volume under the condition of ensuring the data security. For example, the original 512 × 64 pixel image in the first feature sequence is converted into 256 × 128 pixel image or data unit after compression operation, and then at least one operation of extracting the sequence feature and the internal feature is performed, so that the operation cup for extracting the sequence feature and the internal feature is effectively reduced due to the reduction of the data amount. In addition, the compression unit can also perform a second compression operation (encoding), for example, after the image or data unit with 256 × 128 pixels is compressed and converted into the image or data unit with 128 × 256 pixels, then at least one operation of extracting the sequence feature and the internal feature is performed, and because the data amount is reduced again, the operation amount in the operation process is reduced, so that the operation efficiency is improved. In addition, in one embodiment, the compression operation (encoding) of the compression unit may operate through a Max-pooling layer (Max-pooling layer).
The extraction unit is further configured to perform extraction operations of the sequence features and the internal features on the first feature sequence after each compression operation performed by the compression unit. The compressing unit may perform a predetermined number of compression operations on the first feature sequence multiple times according to actual conditions, and the extracting unit performs a predetermined number of operations for extracting the sequence features and the internal features after performing each compression operation. Therefore, the extraction operation of the sequence features and the internal features is carried out for a plurality of times in the whole process, and the accuracy of the data of the target object in the data processing process is ensured.
In an embodiment of the present application, the slice image sequence includes a center slice, and the calculation module is further configured to perform an extraction operation of sequence features on a forward subsequence and a backward subsequence of the center slice, respectively, and correlate results of the two operations to form a forward and backward subsequence set. The central slice is a main slice image of the target object and carries core data of the target object. The central slice corresponds to the forward subsequence and the backward subsequence, and the forward subsequence and the backward subsequence can be regarded as wrapping the central slice therein from two sides respectively. In one embodiment, the priority of the slice image closer to the central slice is higher, and the priority of the slice image farther from the central slice is lower, based on which the calculation module can perform the extraction operation of the sequence feature of the forward subsequence and the extraction operation of the sequence feature of the backward subsequence, so that the slice image closer to the core data plays a larger role, and the accuracy in data processing is improved. Further, performing the above steps may be performed using a separate sequence feature extractor. After the extraction operation is performed, the two operation results can be associated to form a forward and backward subsequence set, and then the forward and backward subsequence set can be associated with the center slice again so as to perform unified processing.
In one embodiment of the present application, in conjunction with fig. 4, the calculation module includes an extraction unit and a decompression unit.
The decompression unit is configured to perform at least one decompression operation on the slice image sequence subjected to the extraction operation to restore the slice image sequence. Specifically, the decompression unit may perform one or more extraction operations on the original slice image sequence (i.e., extract sequence features and internal features, and perform multiple compression operations), and then perform corresponding number of decompression operations to restore the slice image sequence. For example, the decompression unit performs a decompression operation, and converts the original 128 × 256 pixel image or data unit into a 256 × 128 pixel image or data unit after the decompression operation; the decompression unit performs decompression operation again, converts the 256 × 128 pixel image into 512 × 64 pixel image or data unit after decompression operation, and finally restores the slice image sequence.
The extraction unit is configured to perform extraction operation of sequence features on the forward and backward subsequence sets and perform extraction operation of internal features on the central slice after each decompression operation performed by the decompression unit. Specifically, after the extraction unit performs the above operation, the data is not abnormal in the reduction process, and the accuracy of the data is ensured. Further, the extracting unit may associate the operated forward and backward subsequence sets with the center slice after each decompression operation by the decompressing unit so as to perform the next extraction operation of the sequence feature and the extraction operation of the internal feature. In addition, the decompression operation may be performed by using a deconvolution operation.
In an embodiment of the application, the acquisition module is further configured to determine a central slice, select a forward subsequence and a backward subsequence of the central slice according to a preset specification according to a morphology of the target object, and maintain sample balance of the slice image sequence. Specifically, the acquiring module may determine the central slice according to the rough shape of the target object or a position point of interest of a user on the target object, and may select the forward subsequence and the backward subsequence of the central slice according to a preset rule, for example, the acquiring module may extract all slice images in which the target object appears during sampling, and randomly extract a part of slice images in the slice images in which the target object does not appear, with the central slice as a reference, so as to ensure sample balance of the slice image sequence.
In one embodiment of the application, the image processing apparatus further comprises a preprocessing module configured to perform image preprocessing on the slice image according to the physical characteristics of the target object to avoid interference of images of other objects, wherein the image preprocessing includes adapting image attributes of the slice image.
The above embodiments are only exemplary embodiments of the present invention, and are not intended to limit the present invention, and the scope of the present invention is defined by the claims. Various modifications and equivalents may be made by those skilled in the art within the spirit and scope of the present invention, and such modifications and equivalents should also be considered as falling within the scope of the present invention.
Claims (6)
1. An image processing method, comprising:
acquiring a plurality of continuous slice images corresponding to a target object to form a slice image sequence; wherein the sequence of slice images includes a center slice;
extracting sequence features of the sequence of slice images and extracting internal features of the slice images, wherein the sequence features are features characterizing association among a plurality of the slice images; wherein the extracting the sequence features of the sequence of slice images comprises: respectively carrying out extraction operation of the sequence characteristics on the forward subsequence and the backward subsequence of the central slice, and correlating the two operation results to form a forward and backward subsequence set;
performing at least one decompression operation on the slice image sequence after the extraction operation is performed to restore the slice image sequence;
After each decompression operation, performing the sequence feature extraction operation on the forward and backward subsequence sets, and performing the internal feature extraction operation on the central slice;
and acquiring a segmentation map of the target object according to the sequence features and the internal features.
2. The method of claim 1, further comprising:
performing an extraction operation of extracting features on the slice image sequence to generate a first feature sequence;
performing at least one compression operation on the first feature sequence to reduce the operation amount when the sequence features and the internal features are extracted;
and after each compression operation, the extraction operation of the sequence features and the internal features is carried out on the first feature sequence.
3. The method of claim 1, the sequence of slice images including a center slice, the acquiring a plurality of successive slice images corresponding to a target object to form a sequence of slice images including:
and determining the central slice, selecting a forward subsequence and a backward subsequence of the central slice according to a preset specification according to the shape of a target object, and keeping the sample balance of the slice image sequence.
4. The method of claim 1, further comprising:
and performing image preprocessing on the slice image according to the physical characteristics of the target object so as to avoid the interference of the images of other objects, wherein the image preprocessing comprises the step of adaptively changing the image attributes of the slice image.
5. An image processing apparatus includes an acquisition module, a calculation module, and a generation module:
the acquisition module is configured to acquire a plurality of continuous slice images corresponding to a target object to form a slice image sequence; wherein the sequence of slice images includes a center slice;
the computing module is configured to extract sequence features of the sequence of slice images and extract internal features of the slice images, wherein the sequence features are features characterizing associations between a plurality of the slice images; the computing module is further configured to perform the sequence feature extraction operation on the forward subsequence and the backward subsequence of the center slice, and correlate the two operation results to form a forward and backward subsequence set;
the calculation module comprises an extraction unit and a decompression unit;
the decompression unit is configured to perform at least one decompression operation on the slice image sequence after the extraction operation is performed so as to restore the slice image sequence;
The extraction unit is configured to perform the sequence feature extraction operation on the forward and backward subsequence sets and perform the internal feature extraction operation on the central slice after the decompression unit performs each decompression operation;
the generation module is configured to obtain a segmentation map of the target object from the sequence features and the internal features.
6. The apparatus of claim 5, the computation module comprising an extraction unit and a compression unit;
the extraction unit is configured to perform extraction operation of extracting features on the slice image sequence to generate a first feature sequence;
the compression unit is configured to perform at least one compression operation on the first feature sequence to reduce an operation amount when the extraction unit extracts the sequence feature and the internal feature;
the extraction unit is further configured to perform the extraction operation of the sequence feature and the internal feature on the first feature sequence after the compression unit performs the compression operation each time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810437266.0A CN108648195B (en) | 2018-05-09 | 2018-05-09 | Image processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810437266.0A CN108648195B (en) | 2018-05-09 | 2018-05-09 | Image processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108648195A CN108648195A (en) | 2018-10-12 |
| CN108648195B true CN108648195B (en) | 2022-06-28 |
Family
ID=63754042
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810437266.0A Active CN108648195B (en) | 2018-05-09 | 2018-05-09 | Image processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108648195B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2549437A2 (en) * | 2011-07-19 | 2013-01-23 | Kabushiki Kaisha Toshiba | Moving object contour extraction apparatus, left ventricle image segmentation apparatus, moving object contour extraction method and left ventricle image segmentation |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7356190B2 (en) * | 2002-07-02 | 2008-04-08 | Canon Kabushiki Kaisha | Image area extraction method, image reconstruction method using the extraction result and apparatus thereof |
| US20040022447A1 (en) * | 2002-07-31 | 2004-02-05 | General Electric Company | Method and system for image compression and decompression using span of interest of an imaging sequence |
| CN100481112C (en) * | 2007-06-11 | 2009-04-22 | 湖北东润科技有限公司 | Human face recognition system and recognition method |
| CN101610422A (en) * | 2009-07-21 | 2009-12-23 | 天津大学 | Method for compressing three-dimensional image video sequence |
| US8432965B2 (en) * | 2010-05-25 | 2013-04-30 | Intellectual Ventures Fund 83 Llc | Efficient method for assembling key video snippets to form a video summary |
| CN102289811A (en) * | 2011-08-17 | 2011-12-21 | 深圳市旭东数字医学影像技术有限公司 | Semi-automatic sequence image segmentation method and system |
| CN105139377B (en) * | 2015-07-24 | 2017-09-05 | 中南大学 | A kind of fast robust automatic division method of abdominal CT sequence image liver |
| CN106650813B (en) * | 2016-12-27 | 2019-11-15 | 华南理工大学 | A kind of image understanding method based on depth residual error network and LSTM |
| CN107527059B (en) * | 2017-08-07 | 2021-12-21 | 北京小米移动软件有限公司 | Character recognition method and device and terminal |
| CN107967688B (en) * | 2017-12-21 | 2021-06-15 | 联想(北京)有限公司 | Method and system for segmenting object in image |
| CN110569795B (en) * | 2018-03-13 | 2022-10-14 | 腾讯科技（深圳）有限公司 | Image identification method and device and related equipment |
| CN109977816B (en) * | 2019-03-13 | 2021-05-18 | 联想(北京)有限公司 | Information processing method, device, terminal and storage medium |
| CN110363776B (en) * | 2019-06-28 | 2021-10-22 | 联想(北京)有限公司 | Image processing method and electronic device |
-
2018
- 2018-05-09 CN CN201810437266.0A patent/CN108648195B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2549437A2 (en) * | 2011-07-19 | 2013-01-23 | Kabushiki Kaisha Toshiba | Moving object contour extraction apparatus, left ventricle image segmentation apparatus, moving object contour extraction method and left ventricle image segmentation |
Non-Patent Citations (2)
| Title |
|---|
| An Object of Interest based Segmentation Approach for Selective Compression of Video Frames;Marykutty Cyriac等;《Image, Graphics and Signal Processing》;20160229(第2期);第37-44页 * |
| 复杂场景下目标对象的分割与跟踪;朱青祥;《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》;20130715(第07期);第I138-1106页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108648195A (en) | 2018-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110033410B (en) | Image reconstruction model training method, image super-resolution reconstruction method and device | |
| CN111047516B (en) | Image processing method, image processing device, computer equipment and storage medium | |
| CN112200750B (en) | Ultrasonic image denoising model establishing method and ultrasonic image denoising method | |
| CN111179177B (en) | Image reconstruction model training method, image reconstruction method, device and medium | |
| CN111369440B (en) | Model training and image super-resolution processing method, device, terminal and storage medium | |
| CN107730451B (en) | Compressed sensing reconstruction method and system based on depth residual error network | |
| CN108765414B (en) | No-reference stereo image quality evaluation method based on wavelet decomposition and natural scene statistics | |
| CN110689599B (en) | 3D visual saliency prediction method based on non-local enhancement generation countermeasure network | |
| CN106228528B (en) | A kind of multi-focus image fusing method based on decision diagram and rarefaction representation | |
| CN111080567A (en) | Remote sensing image fusion method and system based on multi-scale dynamic convolution neural network | |
| CN111709900A (en) | High dynamic range image reconstruction method based on global feature guidance | |
| CN113436076B (en) | Image super-resolution reconstruction method with characteristics gradually fused and electronic equipment | |
| CN112465801B (en) | Instance segmentation method for extracting mask features in scale division mode | |
| CN110009656B (en) | Target object determination method and device, storage medium and electronic device | |
| CN111178499B (en) | Medical image super-resolution method based on generation countermeasure network improvement | |
| CN113298718A (en) | Single image super-resolution reconstruction method and system | |
| CN112950596A (en) | Tone mapping omnidirectional image quality evaluation method based on multi-region and multi-layer | |
| CN111488810A (en) | Face recognition method and device, terminal equipment and computer readable medium | |
| CN111062953A (en) | Method for identifying parathyroid hyperplasia in ultrasonic image | |
| CN116485741A (en) | No-reference image quality evaluation method, system, electronic equipment and storage medium | |
| CN114418987B (en) | Retina blood vessel segmentation method and system with multi-stage feature fusion | |
| CN114372962A (en) | Laparoscopic surgery stage identification method and system based on double-particle time convolution | |
| Liu et al. | Facial image inpainting using multi-level generative network | |
| CN108648195B (en) | Image processing method and device | |
| CN117315735A (en) | Face super-resolution reconstruction method based on priori information and attention mechanism |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |