CN108712630A - A kind of internet camera system and its implementation based on deep learning - Google Patents
A kind of internet camera system and its implementation based on deep learning Download PDFInfo
- Publication number
- CN108712630A CN108712630A CN201810355057.1A CN201810355057A CN108712630A CN 108712630 A CN108712630 A CN 108712630A CN 201810355057 A CN201810355057 A CN 201810355057A CN 108712630 A CN108712630 A CN 108712630A
- Authority
- CN
- China
- Prior art keywords
- chip
- deep learning
- video front
- image
- artificial intelligence
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
-
- 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
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of internet camera system and its implementation based on deep learning, system includes video front chip and artificial intelligence chip；Method includes：Original image information is pre-processed by video front chip, and pretreated image information is sent to by artificial intelligence chip by USB interface；According to deep learning network model, intellectual analysis is carried out to pretreated image information by artificial intelligence chip；According to intellectual analysis as a result, carrying out subsequent processing by video front chip.The artificial intelligence chip that the present invention increases newly can be connect by USB interface with traditional video front processing chip, and system cost is low；In addition, the present invention carries out intellectual analysis by artificial intelligence chip, the load of server end and the memory bandwidth occupancy of video front are reduced, can be widely applied to internet camera shooting product scope.
Description
Technical field
Image product scope the present invention relates to internet, especially a kind of internet camera system based on deep learning and
Its implementation.
Background technology
Internet camera shooting product is the product that tradition camera shooting product is combined with network video technique, it will be regarded by network
The video data transmission that frequency front-end collection arrives is preserved to the analysis of network distal end.The video of existing most of internet camera shooting products
It is as shown in Fig. 1 to acquire transfer process, Internet video front end sensors acquisition external image data are simultaneously saved in memory, then
Image enhancement is carried out to the image data in memory, video coding technique is then used to carry out compressed encoding to image data, and
Video data is saved as, finally by network transmission to far-end server.However such internet camera shooting product lacks necessity
Intellectual analysis function, the effective information in video data is difficult to obtain timely analyzing processing so that a large amount of invalid videos
Data occupy the memory space of server mass.
The artificial intelligence analysis's function of being now based on deep learning nerual network technique starts to be applied to internet camera shooting production
In product.Compared to pervious image recognition algorithm, the intelligent analysis process based on deep learning network technology has higher identification
Success rate, better fault-tolerance, can meet or exceed the analysis level of the mankind, while calculation amount is also huger.Currently, being
Internet camera shooting product, which increases artificial intelligence analysis's function, mainly following two implementation methods：
(1) increase artificial intelligence analysis's function in server end, that is, after waiting for video data transmission to server, by servicing
Device carries out the artificial intelligence analysis of data to it.Since internet camera shooting product front end is at every moment all carrying out video data
The quantity of collecting work, video data is quite big, and the server of multiple internet camera shooting products front end is connected particularly with those
For, network transmission expense is very huge；And in artificial intelligence analysis's stage of video data, many-one (multiple data origin,
One server) tupe also considerably increase the work load of server, the performance of intellectual analysis is poor.
(2) in internet, artificial intelligence analysis's function is added in camera shooting product front-end chip.This method compared with method (1),
Solve the problems, such as that network overhead is big and server work load is big.However, artificial intelligence analysis's algorithm is computationally intensive, to memory
Requirements for access is huge, and the processing of existing video front has occupied most of system resources in computation and memory bandwidth resource,
Artificial intelligence analysis's function, which is added, will further expand system resource requirements, increase system cost of implementation.
Invention content
In order to solve the above technical problems, it is an object of the invention to：There is provided a kind of load that can reduce server end and
The memory bandwidth occupancy and system cost of video front are low, the internet camera system based on deep learning and its realization
Method.
The first technical solution for being taken of the present invention is：
A kind of internet camera system based on deep learning, including video front chip and artificial intelligence chip,
Wherein, the video front chip, for being pre-processed, being encoded and being stored to original image information；
The artificial intelligence chip, for carrying out intellectual analysis to image using deep learning method；
The video front chip also uploads the result of intellectual analysis；
The video front chip is connect by USB interface with artificial intelligence chip.
Further, the quantity of the artificial intelligence chip is one or more.
Further, the video front chip includes：
Then image processing module carries out post processing of image for carrying out image pre-treatment to original image information；
Microprocessor, for being split processing to the image that post processing of image obtains；
Coding module, the subgraph for being obtained to dividing processing carry out coded treatment；
First memory is stored for the handling result to coding module；
First USB controller, for the handling result of coding module to be transferred to artificial intelligence chip；
Network interface, for realizing the data communication of video front chip and external equipment；
First USB controller is connect with artificial intelligence chip.
Further, the artificial intelligence chip includes the second USB controller, and second USB controller, being used for will be artificial
Intellectual analysis result in intelligent chip is transferred to video front chip, and second USB controller connects with video front chip
It connects.
Further, the artificial intelligence chip further includes：
Decoder module, the data for being sent to video front chip are decoded processing；
Deep learning network module, for according to deep learning network model, intelligence point to be carried out to the result of decoding process
Analysis；
Second memory is stored for the intellectual analysis result to deep learning network module.
The second technical solution for being taken of the present invention is：
A kind of implementation method of the internet camera system based on deep learning, includes the following steps：
Original image information is pre-processed by video front chip, and by USB interface by pretreated figure
As information is sent to artificial intelligence chip；
According to deep learning network model, intelligence point is carried out to pretreated image information by artificial intelligence chip
Analysis, the intellectual analysis includes recognition of face；
According to intellectual analysis as a result, carrying out subsequent processing by video front chip, the subsequent processing includes by intelligence
It can analysis result upload.
Further, the quantity of the artificial intelligence chip is one or more.
Further, described the step for original image information is pre-processed by video front chip, including it is following
Step：
Image pre-treatment is carried out to original image information, then carries out post processing of image；
Processing is split to the image that post processing of image obtains；
Coded treatment is carried out to the subgraph that dividing processing obtains.
Further, described the step for processing is split to the image that post processing of image obtains, specially：
Original image is subjected to longitudinally split or horizontal partition, obtains multiple subgraphs；
Alternatively, extracting multiple boxes for including object from original image, and each box is cut into multiple subgraphs.
Further, it is described according to intellectual analysis as a result, by video front chip carry out subsequent processing the step for, packet
Include following steps：
Integration processing is carried out to intellectual analysis result by video front chip, obtains the intellectual analysis knot of complete image
Fruit；
According to the intellectual analysis of complete image as a result, carrying out subsequent processing by video front chip.
The beneficial effects of the invention are as follows：The artificial intelligence chip that the present invention increases newly by USB interface can with it is traditional
Video front processing chip connects, and artificial intelligence analysis's work(can be had by so that traditional video front chip is not had to redesign
Can, system cost is low；In addition, the present invention carries out intellectual analysis by artificial intelligence chip, it is possible to reduce invalid unrelated video counts
According to network transmission expense, reduce the load of server end and the memory bandwidth occupancy of video front.Further, of the invention
Parallel computation is carried out by multiple artificial intelligence chips, the performance of intellectual analysis can be improved.
Description of the drawings
Fig. 1 is the step flow chart of existing internet camera video acquisition transmission；
Fig. 2 is a kind of overall structure block diagram of the internet camera system based on deep learning of the present invention；
Fig. 3 is that the image block of video front chip transmits schematic diagram；
Fig. 4 is the schematic diagram that internet camera video of the present invention acquires transfer process；
Fig. 5 is a kind of schematic diagram of deep learning network model of the present invention.
Specific implementation mode
The present invention is further explained and is illustrated with specific embodiment with reference to the accompanying drawings of the specification.For of the invention real
The step number in example is applied, is arranged only for the purposes of illustrating explanation, the sequence between step does not do any restriction, implements
The execution sequence of each step in example can be adaptively adjusted according to the understanding of those skilled in the art.
With reference to Fig. 2, a kind of internet camera system based on deep learning, including video front chip and artificial intelligence core
Piece,
Wherein, the video front chip, for being pre-processed, being encoded and being stored to original image information；
The artificial intelligence chip, for carrying out intellectual analysis, the intellectual analysis to image using deep learning method
Including recognition of face；
The video front chip also uploads the result of intellectual analysis；
The video front chip is connect by USB interface with artificial intelligence chip.
It is further used as preferred embodiment, the quantity of the artificial intelligence chip is one or more.
It is further used as preferred embodiment, the video front chip includes：
Then image processing module carries out post processing of image for carrying out image pre-treatment to original image information；
Microprocessor, for being split processing to the image that post processing of image obtains；
Coding module, the subgraph for being obtained to dividing processing carry out coded treatment；
First memory is stored for the handling result to coding module；
First USB controller, for the handling result of coding module to be transferred to artificial intelligence chip；
Network interface, for realizing the data communication of video front chip and external equipment, wherein external equipment includes clothes
Business device, computer and intelligent mobile phone terminal etc.；
First USB controller is connect with artificial intelligence chip.
Wherein, described image processing module, coding module, microprocessor, the first USB controller and network interface are with
One memory connects.
It is further used as preferred embodiment, the artificial intelligence chip includes the second USB controller, and described second
USB controller, for the intellectual analysis result in artificial intelligence chip to be transferred to video front chip, the 2nd USB controls
Device processed is connect with video front chip.
It is further used as preferred embodiment, the artificial intelligence chip further includes：
Decoder module, the data for being sent to video front chip are decoded processing；
Deep learning network module, for according to deep learning network model, intelligence point to be carried out to the result of decoding process
Analysis；
Second memory is stored for the intellectual analysis result to deep learning network module.
Wherein, the decoder module, deep learning network module and the second USB controller are connect with second memory.
Using system as shown in Figure 2, a kind of implementation method of the internet camera system based on deep learning of the invention,
Include the following steps：
Original image information is pre-processed by video front chip, and by USB interface by pretreated figure
As information is sent to artificial intelligence chip；
According to deep learning network model, intelligence point is carried out to pretreated image information by artificial intelligence chip
Analysis, the intellectual analysis includes recognition of face；
According to intellectual analysis as a result, carrying out subsequent processing by video front chip, the subsequent processing includes by intelligence
It can analysis result upload.
It is further used as preferred embodiment, the quantity of the artificial intelligence chip is one or more.
It is further used as preferred embodiment, it is described that original image information is pre-processed by video front chip
The step for, include the following steps：
Image pre-treatment is carried out to original image information, then carries out post processing of image；
Processing is split to the image that post processing of image obtains；
Coded treatment is carried out to the subgraph that dividing processing obtains.
It is further used as preferred embodiment, described be split to the image that post processing of image obtains handles this step
Suddenly, specially：
Original image is subjected to longitudinally split or horizontal partition, obtains multiple subgraphs；
Alternatively, extracting multiple boxes for including object from original image, and each box is cut into multiple subgraphs.
Be further used as preferred embodiment, it is described according to intellectual analysis as a result, being carried out by video front chip
The step for subsequent processing, includes the following steps：
Integration processing is carried out to intellectual analysis result by video front chip, obtains the intellectual analysis knot of complete image
Fruit；
According to the intellectual analysis of complete image as a result, carrying out subsequent processing by video front chip.
The internet camera system of the present invention can be applied to Human detection field, and the person detecting based on high-accuracy is known
Other technology, when detecting someone and entering scene, just take subsequent operation automatically, for example start video record work etc..Below
By taking the application scenarios of recognition of face as an example, a kind of realization of the internet camera system based on deep learning of the present invention is discussed in detail
The specific implementation step of method：
S1, as shown in Fig. 2, video front chip is connected with 5 artificial intelligent chips by USB.When system starts, video
Front-end chip microprocessor reads offline (off-line) trained deep learning network mould from external memory (flash)
Type (the present embodiment use deep learning network model as shown in Figure 5) and model parameter, and by USB by these data transmissions
To 5 artificial intelligent chips；
After S2, artificial intelligence chip receive deep learning network model and the model parameter of the transmission of video front chip,
These data are saved in respective second memory, and incoming deep neural network module；
S3, video front chip are with the speed acquisition image of 25 frames/second and carry out image processing work, described image processing
Work includes pre-treatment and the post-processing etc. to image；
S4, video front chip are split image, are divided into polylith (as shown in figure 3, dividing the image into 5 sons
Image), then image is compressed；
Wherein, step S4 is specially：Original image is subjected to longitudinally split or horizontal partition, obtains multiple subgraphs；
Alternatively, extracting multiple boxes for including object from original image, and each box is cut into multiple subgraphs.
In addition, the present embodiment compresses image by calling JPEG coding modules.
JPEG image data is transferred to multiple artificial intelligence chip (such as Fig. 3 by USB by S5, video front chip respectively
Shown, the present embodiment is equipped with 5 artificial intelligent chips)；
S6, each artificial intelligence chip call JPEG decoder modules to decode image data after receiving jpeg image；
Decoded image data is passed in deep learning network module and handles by S7, artificial intelligence chip.The present embodiment
Using deep learning network model as shown in Figure 5, wherein data indicates that the image data of input, conv indicate convolution operation
(Convolution), relu indicates that line rectification function (Rectified Linear Unit), norm indicate that local acknowledgement is returned
One changes (Local Response Normalization), and pool indicates pond (pooling), and fc indicates full link (fully
Connected), softmax presentation classes function.By using trained model parameter and attached deep learning shown in fig. 5
Network architecture, input picture will obtain intellectual analysis result of calculation after the calculating of a series of function；
Wherein, step S7 specifically includes following steps：
S71, convolutional layer 1：Input data data carries out convolution operation by using 96 3 × 11 × 11 convolution kernels and obtains
96 55 × 55 characteristic patterns, and nonlinear activation is carried out using relu functions, it is normalized using norm；Then, 3 are utilized
× 3 maximum value pond obtains 96 27 × 27 characteristic patterns.
S72, convolutional layer 2：96 × 27 × 27 characteristic pattern carries out convolution behaviour by using 256 96 × 5 × 5 convolution kernels
256 27 × 27 characteristic patterns are obtained, and nonlinear activation is carried out using relu functions, are normalized using norm；It connects
It, 256 13 × 13 characteristic patterns is obtained using 3 × 3 maximum value pond.
S73, convolutional layer 3：256 × 13 × 13 characteristic pattern carries out convolution by using 384 256 × 3 × 3 convolution kernels
Operation obtains 384 13 × 13 characteristic patterns, and carries out nonlinear activation using relu functions.
S74, convolutional layer 4：384 × 13 × 13 characteristic pattern carries out convolution by using 384 384 × 3 × 3 convolution kernels
Operation obtains 384 13 × 13 characteristic patterns, and carries out nonlinear activation using relu functions.
S75, convolutional layer 5：384 × 13 × 13 characteristic pattern carries out convolution by using 256 384 × 3 × 3 convolution kernels
Operation obtains 256 13 × 13 characteristic patterns, and carries out nonlinear activation using relu functions；Then, using 3 × 3 maximum
Value pond obtains 256 6 × 6 characteristic patterns.
S76, full linking layer 1：256 × 6 × 6 characteristic pattern obtains 1 × 4096 characteristic pattern by full linked operation, and makes
Nonlinear activation is carried out with relu functions.
S77, full linking layer 2：1 × 4096 characteristic pattern obtains 1 × 4096 characteristic pattern by full linked operation, and uses
Relu functions carry out nonlinear activation.
S78, full linking layer 3：1 × 4096 characteristic pattern obtains 1 × 1000 characteristic pattern by full linked operation.
S79,1 × 1000 characteristic pattern is converted by intellectual analysis classification results by calling softmax functions.
S8, artificial intelligence chip give analysis result to video front chip by USB transmission；
After S9, video front chip receive each artificial intelligence chip to the analysis result of subgraph, carry out comprehensive whole
Reason, and obtain artificial intelligence analysis's result of whole image；
S10, video front chip microprocessor are according to artificial intelligence analysis as a result, judging whether that someone enters camera shooting field
Scape starts video record work if someone enters；If personage is not present in scene, stop video record work.
S11, the operation processing for repeating step S3-S10, until stopping video record work.Wherein, the tool of S3-S10
Body process flow is as shown in Figure 4.
In conclusion the present invention a kind of internet camera system and its implementation based on deep learning are with following excellent
Point：
(1) present invention before Video coding is transferred to server just progress artificial intelligence analysis, it is possible to reduce video data
Network transmission expense, also reduce the video memory demand of server.
(2) present invention carries out artificial intelligence analysis in video front, it is possible to reduce carries out artificial intelligence point in server end
The calculation amount of analysis can more largely reduce analysis of the server end to multi-channel video when server connects multiple video fronts
Calculation amount.
(3) The present invention reduces the workloads of video front processing chip, since artificial intelligence analysis is operated in manually
It is completed on intelligent chip, avoids the resource that neural computing occupies video front chip, ensured image procossing and video
System resource needed for coding.
(4) present invention has increased artificial intelligence chip newly, can be connected with traditional video front chip by USB interface,
Traditional video front chip is set not have to redesign the function that can have artificial intelligence analysis.
(5) autgmentability of the present invention is strong, can be by multiple artificial intelligence chip parallel computations, without changing chip design
Artificial intelligence analysis's performance of raising system.
It is to be illustrated to the preferable implementation of the present invention, but the present invention is not limited to the embodiment above, it is ripe
Various equivalent variations or replacement can also be made under the premise of without prejudice to spirit of that invention by knowing those skilled in the art, this
Equivalent deformation or replacement are all contained in the application claim limited range a bit.
Claims (10)
1. a kind of internet camera system based on deep learning, it is characterised in that：Including video front chip and artificial intelligence
Chip,
Wherein, the video front chip, for being pre-processed, being encoded and being stored to original image information；
The artificial intelligence chip, for carrying out intellectual analysis to image using deep learning method；
The video front chip also uploads the result of intellectual analysis；
The video front chip is connect by USB interface with artificial intelligence chip.
2. a kind of internet camera system based on deep learning according to claim 1, it is characterised in that：It is described artificial
The quantity of intelligent chip is one or more.
3. a kind of internet camera system based on deep learning according to claim 1, it is characterised in that：The video
Front-end chip includes：
Then image processing module carries out post processing of image for carrying out image pre-treatment to original image information；
Microprocessor, for being split processing to the image that post processing of image obtains；
Coding module, the subgraph for being obtained to dividing processing carry out coded treatment；
First memory is stored for the handling result to coding module；
First USB controller, for the handling result of coding module to be transferred to artificial intelligence chip；
Network interface, for realizing the data communication of video front chip and external equipment；
First USB controller is connect with artificial intelligence chip.
4. a kind of internet camera system based on deep learning according to claim 1, it is characterised in that：It is described artificial
Intelligent chip includes the second USB controller, and second USB controller is used for the intellectual analysis knot in artificial intelligence chip
Fruit is transferred to video front chip, and second USB controller is connect with video front chip.
5. a kind of internet camera system based on deep learning according to claim 4, it is characterised in that：It is described artificial
Intelligent chip further includes：
Decoder module, the data for being sent to video front chip are decoded processing；
Deep learning network module, for according to deep learning network model, intellectual analysis to be carried out to the result of decoding process；
Second memory is stored for the intellectual analysis result to deep learning network module.
6. a kind of implementation method of the internet camera system based on deep learning, it is characterised in that：Include the following steps：
Original image information is pre-processed by video front chip, and is believed pretreated image by USB interface
Breath is sent to artificial intelligence chip；
According to deep learning network model, intellectual analysis, institute are carried out to pretreated image information by artificial intelligence chip
It includes recognition of face to state intellectual analysis；
According to intellectual analysis as a result, carrying out subsequent processing by video front chip, the subsequent processing includes by intelligence point
Result is analysed to upload.
7. a kind of implementation method of internet camera system based on deep learning according to claim 6, feature exist
In：The quantity of the artificial intelligence chip is one or more.
8. a kind of implementation method of internet camera system based on deep learning according to claim 6, feature exist
In：Described the step for original image information is pre-processed by video front chip, include the following steps：
Image pre-treatment is carried out to original image information, then carries out post processing of image；
Processing is split to the image that post processing of image obtains；
Coded treatment is carried out to the subgraph that dividing processing obtains.
9. a kind of implementation method of internet camera system based on deep learning according to claim 8, feature exist
In：Described the step for processing is split to the image that post processing of image obtains, specially：
Original image is subjected to longitudinally split or horizontal partition, obtains multiple subgraphs；
Alternatively, extracting multiple boxes for including object from original image, and each box is cut into multiple subgraphs.
10. a kind of implementation method of internet camera system based on deep learning according to claim 6, feature exist
In：It is described according to intellectual analysis as a result, by video front chip carry out subsequent processing the step for, include the following steps：
Integration processing is carried out to intellectual analysis result by video front chip, obtains the intellectual analysis result of complete image；
According to the intellectual analysis of complete image as a result, carrying out subsequent processing by video front chip.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810355057.1A CN108712630A (en) | 2018-04-19 | 2018-04-19 | A kind of internet camera system and its implementation based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810355057.1A CN108712630A (en) | 2018-04-19 | 2018-04-19 | A kind of internet camera system and its implementation based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108712630A true CN108712630A (en) | 2018-10-26 |
Family
ID=63867303
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810355057.1A Pending CN108712630A (en) | 2018-04-19 | 2018-04-19 | A kind of internet camera system and its implementation based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108712630A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544440A (en) * | 2018-12-29 | 2019-03-29 | 北京知存科技有限公司 | A kind of picture processing chip, System and method for |
| CN110264655A (en) * | 2019-06-24 | 2019-09-20 | 天津华来科技有限公司 | Intelligent monitoring device and system |
| CN110493598A (en) * | 2019-08-12 | 2019-11-22 | 北京中科寒武纪科技有限公司 | Method for processing video frequency and relevant apparatus |
| CN111242321A (en) * | 2019-04-18 | 2020-06-05 | 中科寒武纪科技股份有限公司 | Data processing method and related product |
| CN114255198A (en) * | 2019-03-28 | 2022-03-29 | 芨影(厦门)科技有限公司 | Artificial intelligence modeling and identification system for medical graphic image AI |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120275690A1 (en) * | 2011-04-26 | 2012-11-01 | Nec Laboratories America, Inc. | Distributed artificial intelligence services on a cell phone |
| US20140300758A1 (en) * | 2013-04-04 | 2014-10-09 | Bao Tran | Video processing systems and methods |
| CN104320615A (en) * | 2014-10-17 | 2015-01-28 | 智擎信息系统（上海）有限公司 | Intelligent video security and protection system and signal processing method thereof |
| CN105893159A (en) * | 2016-06-21 | 2016-08-24 | 北京百度网讯科技有限公司 | Data processing method and device |
| CN205621018U (en) * | 2016-02-26 | 2016-10-05 | 陈进民 | Cell -phone cell convolutional neural network accelerator |
| CN205621017U (en) * | 2016-02-22 | 2016-10-05 | 陈进民 | Intelligent vision chip |
| CN205987179U (en) * | 2016-08-29 | 2017-02-22 | 内蒙古智诚物联股份有限公司 | Video flowing artificial intelligence analysis module |
| CN106897695A (en) * | 2017-02-24 | 2017-06-27 | 上海斐讯数据通信技术有限公司 | A kind of image recognizing and processing equipment, system and method |
| CN107067365A (en) * | 2017-04-25 | 2017-08-18 | 中国石油大学(华东) | The embedded real-time video stream processing system of distribution and method based on deep learning |
| CN206648005U (en) * | 2017-04-19 | 2017-11-17 | 杭州派尼澳电子科技有限公司 | A kind of intelligent road-lamp of built-in impact sound detection function |
| CN107808389A (en) * | 2017-10-24 | 2018-03-16 | 上海交通大学 | Unsupervised methods of video segmentation based on deep learning |
| CN107832841A (en) * | 2017-11-14 | 2018-03-23 | 福州瑞芯微电子股份有限公司 | The power consumption optimization method and circuit of a kind of neural network chip |
-
2018
- 2018-04-19 CN CN201810355057.1A patent/CN108712630A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120275690A1 (en) * | 2011-04-26 | 2012-11-01 | Nec Laboratories America, Inc. | Distributed artificial intelligence services on a cell phone |
| US20140300758A1 (en) * | 2013-04-04 | 2014-10-09 | Bao Tran | Video processing systems and methods |
| CN104320615A (en) * | 2014-10-17 | 2015-01-28 | 智擎信息系统（上海）有限公司 | Intelligent video security and protection system and signal processing method thereof |
| CN205621017U (en) * | 2016-02-22 | 2016-10-05 | 陈进民 | Intelligent vision chip |
| CN205621018U (en) * | 2016-02-26 | 2016-10-05 | 陈进民 | Cell -phone cell convolutional neural network accelerator |
| CN105893159A (en) * | 2016-06-21 | 2016-08-24 | 北京百度网讯科技有限公司 | Data processing method and device |
| CN205987179U (en) * | 2016-08-29 | 2017-02-22 | 内蒙古智诚物联股份有限公司 | Video flowing artificial intelligence analysis module |
| CN106897695A (en) * | 2017-02-24 | 2017-06-27 | 上海斐讯数据通信技术有限公司 | A kind of image recognizing and processing equipment, system and method |
| CN206648005U (en) * | 2017-04-19 | 2017-11-17 | 杭州派尼澳电子科技有限公司 | A kind of intelligent road-lamp of built-in impact sound detection function |
| CN107067365A (en) * | 2017-04-25 | 2017-08-18 | 中国石油大学(华东) | The embedded real-time video stream processing system of distribution and method based on deep learning |
| CN107808389A (en) * | 2017-10-24 | 2018-03-16 | 上海交通大学 | Unsupervised methods of video segmentation based on deep learning |
| CN107832841A (en) * | 2017-11-14 | 2018-03-23 | 福州瑞芯微电子股份有限公司 | The power consumption optimization method and circuit of a kind of neural network chip |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544440A (en) * | 2018-12-29 | 2019-03-29 | 北京知存科技有限公司 | A kind of picture processing chip, System and method for |
| CN114255198A (en) * | 2019-03-28 | 2022-03-29 | 芨影(厦门)科技有限公司 | Artificial intelligence modeling and identification system for medical graphic image AI |
| CN111242321A (en) * | 2019-04-18 | 2020-06-05 | 中科寒武纪科技股份有限公司 | Data processing method and related product |
| US11762690B2 (en) | 2019-04-18 | 2023-09-19 | Cambricon Technologies Corporation Limited | Data processing method and related products |
| CN111242321B (en) * | 2019-04-18 | 2023-09-26 | 中科寒武纪科技股份有限公司 | Data processing method and related product |
| CN110264655A (en) * | 2019-06-24 | 2019-09-20 | 天津华来科技有限公司 | Intelligent monitoring device and system |
| CN110493598A (en) * | 2019-08-12 | 2019-11-22 | 北京中科寒武纪科技有限公司 | Method for processing video frequency and relevant apparatus |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108712630A (en) | A kind of internet camera system and its implementation based on deep learning | |
| CN108629743A (en) | Processing method, device, storage medium and the electronic device of image | |
| CN111479048A (en) | Intelligent video image processing equipment based on edge calculation | |
| WO2023174098A1 (en) | Real-time gesture detection method and apparatus | |
| CN110033083A (en) | Convolutional neural networks model compression method and apparatus, storage medium and electronic device | |
| WO2021155832A1 (en) | Image processing method and related device | |
| CN109711407B (en) | License plate recognition method and related device | |
| CN110099280B (en) | Video service quality enhancement method under limitation of wireless self-organizing network bandwidth | |
| CN111510774A (en) | Intelligent terminal image compression algorithm combining edge calculation and deep learning | |
| CN110298296B (en) | Face recognition method applied to edge computing equipment | |
| CN106254458A (en) | A kind of image processing method based on cloud robot vision, platform and system | |
| CN106897695A (en) | A kind of image recognizing and processing equipment, system and method | |
| CN107133590A (en) | A kind of identification system based on facial image | |
| WO2023005740A1 (en) | Image encoding, decoding, reconstruction, and analysis methods, system, and electronic device | |
| CN112446244A (en) | Human body action recognition method, neural network training method, related device and equipment | |
| CN111209119A (en) | Load balancing method for face snapshot rifle bolt | |
| CN114360018A (en) | Rendering method and device of three-dimensional facial expression, storage medium and electronic device | |
| CN116524195B (en) | Semantic segmentation method, semantic segmentation device, electronic equipment and storage medium | |
| CN113705384A (en) | Facial expression recognition method considering local space-time characteristics and global time sequence clues | |
| WO2023174256A1 (en) | Data compression method and related device | |
| CN107770539A (en) | The coding method of character block and device and its coding/decoding method and device | |
| CN112070797A (en) | Target detection method, system, acceleration device, medium and electronic equipment | |
| WO2019127926A1 (en) | Calculation method and calculation device for sparse neural network, electronic device, computer readable storage medium, and computer program product | |
| CN112070211B (en) | Image recognition method based on computing unloading mechanism | |
| WO2022022176A1 (en) | Image processing method and related device |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181026 |
|
| RJ01 | Rejection of invention patent application after publication |