CN113052751B - Artificial intelligence processing method - Google Patents
Artificial intelligence processing method Download PDFInfo
- Publication number
- CN113052751B CN113052751B CN202110461703.4A CN202110461703A CN113052751B CN 113052751 B CN113052751 B CN 113052751B CN 202110461703 A CN202110461703 A CN 202110461703A CN 113052751 B CN113052751 B CN 113052751B
- Authority
- CN
- China
- Prior art keywords
- data
- forwarding
- unit
- area
- processor
- 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
- G06T1/00—General purpose image data processing
- G06T1/20—Processor architectures; Processor configuration, e.g. pipelining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/60—Memory management
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The application provides an artificial intelligence processing method, which is applied to a terminal, wherein the terminal comprises the following steps: the device comprises a processor, a memory, a camera and a display screen; the processor includes: the system comprises a general processor, an image processor GPU and an AI processor; wherein the general-purpose processor, the image processor, and the AI processor are connected to each other. The technical scheme provided by the application has the advantage of high video processing speed.
Description
Technical Field
The application relates to the technical field of image processing, in particular to an artificial intelligence processing method.
Background
Graphics processors (English: graphics Processing Unit, abbreviated: GPU), also known as display cores, vision processors, display chips, are microprocessors that are dedicated to performing image and graphics related operations on personal computers, workstations, gaming machines, and some mobile devices (e.g., tablet computers, smartphones, etc.).
The existing AI robot virtualized by the GPU influences the processing speed of the video.
Disclosure of Invention
The embodiment of the application provides an artificial intelligence processing method which has the advantage of high image processing speed.
In a first aspect, an embodiment of the present application provides an artificial intelligence processing method, where the method is applied to a terminal, and the terminal includes: the device comprises a processor, a memory, a camera and a display screen; the processor includes: the system comprises a general processor, an image processor GPU and an AI processor; wherein the general-purpose processor, the image processor and the AI processor are connected with each other; the method comprises the following steps:
the general processor reads the front x frame data of the video from the memory and sends the front x frame data to the GPU;
the GPU performs image processing on the front x frame data to obtain front x frame picture data; each frame of picture of the previous x frame of picture data is sent to an AI processor;
the AI processor identifies and determines a target object area and a background area of each frame of picture, and superimposes the first x target object areas of the previous x frame of picture data to obtain a range of a first area of the video and a range of a background area of the video; transmitting the range of the first region to a general purpose processor;
the general processor determines the subsequent data of the video x frame, and sends partial data corresponding to the range of the first area in the subsequent data to the GPU;
and the GPU calculates the partial data to obtain partial picture data of the follow-up data, and splices the partial picture data with the data corresponding to the range of the background area of the video to obtain follow-up picture result data.
The implementation of the embodiment of the application has the following beneficial effects:
it can be seen that when the technical scheme provided by the application processes the video, the first x frames of pictures of the video are processed completely, then the subsequent data are processed only by the partial data corresponding to the range of the first area, and the pictures of the subsequent data are obtained after simple splicing, so that the processing capacity of the subsequent data is reduced, the video processing efficiency is improved, the operation capacity is reduced, and the processing speed of the video is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic structure of a terminal.
Fig. 2 is a schematic structural diagram of a processor according to an embodiment of the present application.
Fig. 3 is a schematic structural diagram of the GPU provided in the present application.
Fig. 4 is a schematic distribution diagram of a signaling sub-region, a data sub-region, and a composite sub-region provided in the present application.
Fig. 5 is a schematic structural diagram of an AI processor provided in the present application.
Fig. 6 is a schematic diagram of the regions of two frames and a target object of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.
The terms "first," "second," "third," and "fourth" and the like in the description and in the claims of this application and in the drawings, are used for distinguishing between different objects and not for describing a particular sequential order. Furthermore, the terms "comprise" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, result, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.
Referring to fig. 1, fig. 1 provides a terminal (may also be referred to as a GPU-based virtualized AI robot management device or system), which may specifically include: the processor, the memory, the camera and the display screen can be connected through a bus, or can be connected through other modes, and the specific mode of connection is not limited in the application.
Referring to fig. 2, the processor may specifically include: general purpose processor, image processor GPU and AI processor. The artificial intelligence processing method comprises the following steps: service robot, intelligent fitting equipment.
The general processor, the image processor GPU and the AI processor are connected with each other in pairs.
Referring to fig. 3, fig. 3 provides a structure of a GPU, referring to fig. 3 (taking m=6, n=6 as an example, where m may be actually greater than 6, e.g. 12, 18, etc.), fig. 3 provides a structure of a GPU, and the structure may specifically include: a storage unit 201, n control units 202, a calculation unit 203, a signaling forwarding unit 204, a data forwarding unit 205, and a compound forwarding unit 206; the sum of the numbers of the calculation unit 203, the signaling forwarding unit 204, the data forwarding unit 205 and the compound forwarding unit 206 is m×n;
wherein the n control units 202 are arranged in rows, m is a row value of the matrix, n is a column value of the matrix, and the n control units 202 are respectively connected with n units of a first row of the m is a n units of the matrix; the memory cells 201 are connected to m cells of the last row of the matrix arrangement, respectively;
the m×n units arranged in matrix include a normal area 301, a signaling forwarding area 302, a data forwarding area 303, and a composite forwarding area 304; wherein the normal area 301 includes only the calculation unit 203, and the signaling forwarding area 302 includes: a calculation unit and a signaling forwarding unit; the data transfer area 303 includes: the calculation unit and the data forwarding unit, the composite forwarding area 304 includes: a computing unit and a compound forwarding unit;
the control unit is used for sending calculation instructions to the calculation unit, the signaling forwarding unit and the composite forwarding unit; for ease of drawing, the connections of the control unit with the signalling forwarding unit and the compound forwarding unit are indicated with dashed lines,
a storage unit 201 for storing calculation data or calculation results;
the storage unit 201 has a plurality of IO (input output) interfaces, which are respectively connected to m computing units, data forwarding units, and compound forwarding units of the last row of the matrix arrangement; for convenience of drawing, the connection of the storage unit with the data forwarding unit and the composite forwarding unit is indicated by a dotted line,
a calculation unit for performing operations (arithmetic operations such as addition, subtraction, multiplication, division, etc.) on calculation data (calculation data that may be read or received) according to the calculation instruction to obtain a calculation result; the calculation result is sent to a storage unit (if the calculation result is connected with an IO interface of the storage unit, the calculation result is directly sent to the storage unit, if the calculation result is not directly connected with the IO interface of the storage unit, the calculation result is sent to the storage unit in a forwarding mode (the calculation unit can be directly connected with the storage unit through a composite forwarding unit);
the signaling forwarding area comprises a plurality of signaling subareas, each signaling subarea is arranged in a 3*3 array (shown in fig. 3), the signaling forwarding unit is positioned at the center of the 3*3 array, the signaling forwarding units are respectively connected with 8 computing units at the edge of the 3*3 array, and the signaling forwarding units are used for receiving the computing instructions sent by the control unit and forwarding the computing instructions to 8 computing units at the edge of the 3*3 array;
the data forwarding area comprises a plurality of data subareas, each data subarea is arranged in a 3*3 array (as shown in fig. 3), the data forwarding units are positioned at the central position of the 3*3 array, the data forwarding units are respectively connected with 8 computing units at the edge of the 3*3 array, and the data forwarding units are used for extracting the computing data of the storage units and forwarding the computing data to 8 computing units at the edge of the array of the control unit 3*3;
the compound forwarding area comprises a plurality of compound sub-areas, each compound sub-area is in 3*3 array arrangement (as shown in fig. 3), the compound forwarding unit is located at the central position of the 3*3 array, and the compound forwarding units are respectively connected with 8 computing units at the edge of the 3*3 array, and are used for receiving the computing instructions sent by the control unit, extracting the computing data of the storage unit, and forwarding the computing instructions and the computing data to 8 computing units at the edge of the array of the control unit 3*3.
And m and n are integers greater than or equal to 5, and m is greater than or equal to n.
As shown in fig. 4, adjacent computing units are interconnected, which may be for the transfer of data or signaling. The adjacent may be upper and lower, or left and right.
The influence on the calculation speed mainly comprises 2 directions, namely the first direction, namely the calculation speed is high, namely the same data calculation speed is high, the calculation speed is mainly based on the frequency of a processing circuit, the second direction is low in IO overhead, namely the same data forwarding frequency is low, for the GPU structure, because the calculation units are many, if all the calculation units are directly connected with a control unit and a storage unit, the number of interfaces of the control unit and the storage unit can be greatly increased, the cost can be greatly increased, so that the number of IO interfaces is not increased, the forwarding frequency is required to be reduced, 4 areas are divided based on the thought, and different forwarding units (namely forwarding circuits) are respectively arranged for the characteristics of the 4 areas to realize different functions, so that the speed of image processing is improved.
The technical scheme provided by the application has the technical effects that the number of times of forwarding the calculation data or the calculation signaling is reduced, namely, the number of times of forwarding the calculation data or the calculation signaling is reduced by arranging corresponding forwarding units in different areas, the time delay of the calculation data and the calculation signaling is reduced, and the processing speed of the image is further improved.
Each of the above units may be implemented in hardware circuits including, but not limited to, FPGAs, CGRAs, application specific integrated circuits ASICs, analog circuits, memristors, and the like.
Referring to fig. 5, fig. 5 provides a structure of an AI processor for performing a neural network forward operation; the neural network comprises n layers; the AI processor has a structure as shown in fig. 5, and includes:
the AI processor includes: the system comprises a main processing circuit, k branch processing circuits and k groups of basic processing circuits, wherein the main processing circuit is respectively connected with the k branch processing circuits, each branch processing circuit in the k branch processing circuits corresponds to one group of basic processing circuits in the k groups of basic processing circuits, and the group of basic processing circuits comprises at least one basic processing circuit.
Of course, in practical applications, the AI processor may also be a general-purpose AI processor, such as the processing chip of the source 270.
The artificial intelligence processing method as shown in fig. 1 may be specifically used to perform video processing, and specifically may include:
the general processor reads the front x frame data of the video from the memory and sends the front x frame data to the GPU;
the GPU performs image processing on the front x frame data to obtain front x frame picture data; each frame of picture of the previous x frame of picture data is sent to an AI processor;
the implementation manner of performing image processing on the previous x frame data to obtain the previous x frame picture data may be specifically referred to the operation manner of the GPU structure shown in fig. 2 and 3, where the calculation data in the GPU structure shown in fig. 2 and 3 may be the previous x frame data, and the calculation result may be the previous x frame picture data.
Of course, the image processing performed on the previous x frame data to obtain the previous x frame image data may also be processed by adopting a through image processing manner, which is not described herein again.
The AI processor identifies and determines a target object area and a background area of each frame of picture, and superimposes the first x target object areas of the previous x frame of picture data to obtain the range of a first area of the video (namely, the set of the areas where the x target objects possibly appear obtained by identification) and the range of the background area of the video; transmitting the range of the first region to a general purpose processor;
specifically, the range of the first region of the video obtained by overlapping the first x target object regions of the previous x frame of picture data may specifically include:
all the area identifications of the first x target object areas are determined, and the area identifications are collected to be determined as the range of the first area.
For example, the first x target object regions relate to region 3, region 4, and region 5, and then the first region may range from region 3, region 4, and region 5. The area range and the number of the areas contained in the pictures can be set by a manufacturer, and because the background is fixed, the size of each frame of picture acquisition of the video frames is consistent, a rectangular frame is established in the size, and each rectangular frame can represent one area.
The general processor determines the subsequent data of the video x frame, and sends partial data corresponding to the range of the first area in the subsequent data to the GPU;
the subsequent data may be data corresponding to video frames after the x-th frame of the video.
And the GPU calculates the partial data to obtain partial picture data of the follow-up data, and splices the partial picture data with the data corresponding to the range of the background area of the video to obtain follow-up picture result data.
The subsequent picture result data may include: data in different picture formats, such as JPGE, PDF, etc.
The implementation manner of calculating the partial data to obtain the partial picture of the subsequent data may specifically refer to the operation manner of the GPU structure shown in fig. 2 and 3, where the calculated data in the GPU structure shown in fig. 2 and 3 may be the partial data, and the calculation result may be the partial picture.
Of course, the above-mentioned partial image for calculating the partial data to obtain the subsequent data may also be processed by adopting a through image processing manner, which is not described herein again.
The value of x may be a smaller integer, for example, the value of x may be in the range of [ 10,100 ].
When the technical scheme provided by the application is used for processing the video, the front x frames of pictures of the video are processed completely, then the subsequent data are processed only by partial data corresponding to the range of the first area, and the pictures of the subsequent data are obtained by simple splicing, so that the processing capacity of the subsequent data is reduced, the video processing efficiency is improved, the operation capacity is reduced, and the processing speed of the video is improved.
For example, in an alternative solution, the method may further include:
the general processor establishes a mapping relation between each frame of data of the video and the region range in advance, and the mapping relation can be used for determining partial data corresponding to the range of the first region. The mapping relationship may specifically be a storage space set corresponding to the region identifier, the range of the first region may be a region identifier set (the range of the background region may be a region identifier set other than the range of the first region in the video frame), for example, the range of the first region includes a region 3, a region 4, and a region 5, and if it is determined according to the mapping relationship, the storage space 3, the storage space 4, and the storage space 5 corresponding to the region 3, the region 4, and the region 5, it is determined that the portion of data is data stored in the storage space 3, the storage space 4, and the storage space 5.
The video of the present application may specifically be a video with a fixed background, that is, a video in which the angle of the camera will not rotate, for an artificial intelligence processing method, for example, a common service robot in a mall, although it may move, but is generally not moved during a period of time operated by a user, the collected video is a video with a fixed background (because the angle and the position of the image are fixed), for such video, that is, the background is fixed, but the target object (for example, the user) will move, but the area where the video is shot is specific, for all frames of the video, only the area range of the target object in the video frame picture is different, as shown in fig. 6, from frame 1 to frame 2, the target object 601 is actually moved only by a distance x, then, if the mapping relationship between each frame data and the area range of the target object is pre-established, after determining the area range of the target object by the AI processor, only the image processing is required for the data corresponding to the area range of the target object for the subsequent video frame, and the number of times of operations can be reduced by the data corresponding to the area range of the background frame (that is the area corresponding to the previous frame x) of the video.
Referring to fig. 6, the box in fig. 6 is a region, the inside number indicates the identification of each region, only 5 regions are identified here, the dotted line in fig. 6 indicates the range of the target object 601 in frame 1, and the solid line indicates the range of the target object 601 in frame 2.
It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of action combinations, but it should be understood by those skilled in the art that the present application is not limited by the order of actions described, as some steps may be performed in other order or simultaneously in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all alternative embodiments, and that the acts and modules referred to are not necessarily required in the present application.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.
The foregoing has outlined rather broadly the more detailed description of embodiments of the present application, wherein specific examples are provided herein to illustrate the principles and embodiments of the present application, the above examples being provided solely to assist in the understanding of the methods of the present application and the core ideas thereof; meanwhile, as those skilled in the art will have modifications in the specific embodiments and application scope in accordance with the ideas of the present application, the present description should not be construed as limiting the present application in view of the above.
Claims (6)
1. An artificial intelligence processing method, the method being applied to a terminal, the terminal comprising: the device comprises a processor, a memory, a camera and a display screen; the processor includes: the system comprises a general processor, an image processor GPU and an AI processor; wherein the general-purpose processor, the image processor and the AI processor are connected with each other; characterized in that the method comprises:
the general processor reads the front x frame data of the video from the memory and sends the front x frame data to the GPU;
the GPU performs image processing on the front x frame data to obtain front x frame picture data; each frame of picture of the previous x frame of picture data is sent to an AI processor;
the AI processor identifies and determines a target object area and a background area of each frame of picture, and superimposes the first x target object areas of the previous x frame of picture data to obtain a range of a first area of the video and a range of a background area of the video; transmitting the range of the first region to a general purpose processor;
the general processor determines the subsequent data of the video x frame, and sends partial data corresponding to the range of the first area in the subsequent data to the GPU;
the GPU calculates the partial data to obtain partial picture data of the subsequent data, and splices the partial picture data with data corresponding to the range of the background area of the video to obtain subsequent picture result data;
the GPU comprises: the device comprises a storage unit, n control units, a calculation unit, a signaling forwarding unit, a data forwarding unit and a compound forwarding unit; the sum of the numbers of the calculation unit, the signaling forwarding unit, the data forwarding unit and the compound forwarding unit is m x n;
the n control units are arranged in columns, m is an n-unit matrix, wherein m is a row value of the matrix, n is a column value of the matrix, and the n control units are respectively connected with n units of a first column of the m is an n-unit matrix; the storage units are respectively connected with m units of the last row of the matrix arrangement;
the m x n units arranged in the matrix comprise a common area, a signaling forwarding area, a data forwarding area and a composite forwarding area; wherein, the common area only comprises a calculation unit, and the signaling forwarding area comprises: a calculation unit and a signaling forwarding unit; the data forwarding area includes: the data forwarding unit is used for forwarding data in a composite forwarding area, and the composite forwarding area comprises: a computing unit and a compound forwarding unit;
the computing unit is used for executing operation on the computing data according to the computing instruction to obtain a computing result; transmitting the calculation result to a storage unit;
the signaling forwarding area comprises a plurality of signaling subareas, each signaling subarea is arranged in a 3*3 array, the signaling forwarding unit is positioned at the central position of the 3*3 array and is respectively connected with 8 computing units at the edge of the 3*3 array, and the signaling forwarding unit is used for receiving the computing instructions sent by the control unit and forwarding the computing instructions to 8 computing units at the edge of the array of the control unit 3*3;
the data forwarding area comprises a plurality of data subareas, each data subarea is arranged in a 3*3 array, the data forwarding unit is positioned at the center of the 3*3 array and is respectively connected with 8 computing units at the edge of the 3*3 array, and the data forwarding unit is used for extracting the computing data of the storage unit and forwarding the computing data to 8 computing units at the edge of the control unit 3*3 array;
the composite forwarding area comprises a plurality of composite sub-areas, each composite sub-area is arranged in a 3*3 array, the composite forwarding unit is positioned at the center of the 3*3 array and is respectively connected with 8 computing units at the edge of the 3*3 array, and the composite forwarding unit is used for receiving the computing instructions sent by the control unit, extracting the computing data of the storage unit and forwarding the computing instructions and the computing data to 8 computing units at the edge of the array of the control unit 3*3;
and m and n are integers greater than or equal to 5, and m is greater than or equal to n.
2. The artificial intelligence processing method of claim 1, wherein the artificial intelligence processing method comprises the steps of,
the general processor establishes a mapping relation between each frame of data of the video and the region range in advance.
3. The artificial intelligence processing method according to claim 1, wherein the overlapping the first x target object regions of the previous x frame picture data to obtain the first region of the video comprises
All the area identifications of the first x target object areas are determined, and the area identifications are collected to be determined as the range of the first area.
4. A method according to any one of the claims 1-3, characterized in that,
the AI processor includes: the system comprises a main processing circuit, k branch processing circuits and k groups of basic processing circuits, wherein the main processing circuit is respectively connected with the k branch processing circuits, each branch processing circuit in the k branch processing circuits corresponds to one group of basic processing circuits in the k groups of basic processing circuits, and the group of basic processing circuits comprises at least one basic processing circuit.
5. The artificial intelligence processing method according to claim 1, wherein the data forwarding unit and the composite forwarding unit are respectively provided with registers.
6. The artificial intelligence processing method of claim 1, wherein the artificial intelligence processing method comprises the steps of,
the terminal is as follows: service robot, intelligent fitting equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110461703.4A CN113052751B (en) | 2021-04-27 | 2021-04-27 | Artificial intelligence processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110461703.4A CN113052751B (en) | 2021-04-27 | 2021-04-27 | Artificial intelligence processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113052751A CN113052751A (en) | 2021-06-29 |
| CN113052751B true CN113052751B (en) | 2023-05-30 |
Family
ID=76517715
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110461703.4A Active CN113052751B (en) | 2021-04-27 | 2021-04-27 | Artificial intelligence processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113052751B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016086754A1 (en) * | 2014-12-03 | 2016-06-09 | 中国矿业大学 | Large-scale scene video image stitching method |
| CN108519807A (en) * | 2018-03-23 | 2018-09-11 | 维沃移动通信有限公司 | A kind of application processor and mobile terminal |
| JP2019046199A (en) * | 2017-09-01 | 2019-03-22 | 株式会社半導体エネルギー研究所 | Processor and electronic device |
| CN110852942A (en) * | 2019-11-19 | 2020-02-28 | 腾讯科技（深圳）有限公司 | Model training method, and media information synthesis method and device |
| US10825141B1 (en) * | 2019-08-19 | 2020-11-03 | Samsung Electronics Co., Ltd. | Decoding apparatus and operating method of the same, and artificial intelligence (AI) upscaling apparatus and operating method of the same |
| CN112183736A (en) * | 2019-07-05 | 2021-01-05 | 三星电子株式会社 | Artificial intelligence processor and method for executing neural network operation |
| CN212727148U (en) * | 2020-06-29 | 2021-03-16 | 北京百度网讯科技有限公司 | Artificial intelligence AI camera |
-
2021
- 2021-04-27 CN CN202110461703.4A patent/CN113052751B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016086754A1 (en) * | 2014-12-03 | 2016-06-09 | 中国矿业大学 | Large-scale scene video image stitching method |
| JP2019046199A (en) * | 2017-09-01 | 2019-03-22 | 株式会社半導体エネルギー研究所 | Processor and electronic device |
| CN108519807A (en) * | 2018-03-23 | 2018-09-11 | 维沃移动通信有限公司 | A kind of application processor and mobile terminal |
| CN112183736A (en) * | 2019-07-05 | 2021-01-05 | 三星电子株式会社 | Artificial intelligence processor and method for executing neural network operation |
| US10825141B1 (en) * | 2019-08-19 | 2020-11-03 | Samsung Electronics Co., Ltd. | Decoding apparatus and operating method of the same, and artificial intelligence (AI) upscaling apparatus and operating method of the same |
| CN110852942A (en) * | 2019-11-19 | 2020-02-28 | 腾讯科技（深圳）有限公司 | Model training method, and media information synthesis method and device |
| CN212727148U (en) * | 2020-06-29 | 2021-03-16 | 北京百度网讯科技有限公司 | Artificial intelligence AI camera |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113052751A (en) | 2021-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110096310B (en) | Operation method, operation device, computer equipment and storage medium | |
| WO2019184657A1 (en) | Image recognition method, apparatus, electronic device and storage medium | |
| WO2019089816A3 (en) | System having a hybrid threading processor, a hybrid threading fabric having configurable computing elements, and a hybrid interconnection network | |
| US20220282993A1 (en) | Map fusion method, device and storage medium | |
| CN109993301B (en) | Neural network training device and related product | |
| CN109359556B (en) | Face detection method and system based on low-power-consumption embedded platform | |
| WO2021036729A1 (en) | Matrix computation method, computation device, and processor | |
| CN112734827A (en) | Target detection method and device, electronic equipment and storage medium | |
| CN110909872A (en) | Integrated circuit chip device and related product | |
| CN113052751B (en) | Artificial intelligence processing method | |
| EP3497667A1 (en) | Apparatus, method, and computer program code for producing composite image | |
| CN113837922A (en) | Computing device, data processing method and related product | |
| CN112084023A (en) | Data parallel processing method, electronic equipment and computer readable storage medium | |
| CN111382847B (en) | Data processing device and related product | |
| CN109740730B (en) | Operation method, device and related product | |
| WO2023184754A1 (en) | Configurable real-time disparity point cloud computing apparatus and method | |
| CN111260046B (en) | Operation method, device and related product | |
| CN114254563A (en) | Data processing method and device, electronic equipment and storage medium | |
| CN112308787B (en) | Distortion correction method and device and electronic equipment | |
| CN111260070B (en) | Operation method, device and related product | |
| CN111258641B (en) | Operation method, device and related product | |
| CN114332928A (en) | Image segmentation method and device, electronic equipment and storage medium | |
| CN112395008A (en) | Operation method, operation device, computer equipment and storage medium | |
| CN114443137A (en) | Integrated computing device, chip, board card, equipment and computing method | |
| CN111061507A (en) | Operation method, operation device, computer equipment and storage medium |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230515 Address after: Room 623, Building 1, No. 132, Shenjia Road, Dongxin Street, Xiacheng District, Hangzhou City, Zhejiang Province, 310006 Applicant after: Zhejiang Shuike Culture Group Co.,Ltd. Address before: 1113, qiurui building, 207 Minkang Road, Zhangkeng community, Minzhi street, Longhua District, Shenzhen, Guangdong 518000 Applicant before: Shenzhen Dongfang Maizhuo Technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |