CN108462876B - Video decoding optimization adjustment device and method - Google Patents
Video decoding optimization adjustment device and method Download PDFInfo
- Publication number
- CN108462876B CN108462876B CN201810051581.XA CN201810051581A CN108462876B CN 108462876 B CN108462876 B CN 108462876B CN 201810051581 A CN201810051581 A CN 201810051581A CN 108462876 B CN108462876 B CN 108462876B
- Authority
- CN
- China
- Prior art keywords
- unit
- scene
- image
- content
- neural network
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/20—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/179—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a scene or a shot
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a video decoding optimization adjusting device and a method, wherein in a training mode, a video decoding circuit decodes a video source file, and sends a decoded continuous video image with a label to a deep learning circuit to identify scenes and contents so as to finish training; in an optimization mode, the video decoding circuit sends decoded continuous video images to the deep learning circuit, the deep learning circuit identifies scenes and contents and sends an identification result to the optimization adjusting circuit, and the optimization adjusting circuit carries out optimization adjustment processing on the video images according to the scene and content identification result and then sends the video images to the display control unit for display. The invention can lead the video decoder to learn the current decoding image content and scene classification by the neural network and make different optimization adjustments to the classification result so as to achieve the best decoding effect.
Description
Technical Field
The invention relates to a video decoding optimization adjustment circuit and a video decoding optimization adjustment method, which are based on the video decoding of a video decoding circuit and realize the classification and optimization adjustment of contents or scenes through deep learning of video contents.
Background
The current video decoding and playing are to directly play the decoded content to the user, and cannot be optimized and adjusted according to the current playing content. If the video decoder can learn and recognize the current decoded image content and scene classification, and make different decoding adjustments according to the scene classification, the video playing effect and the user experience can be greatly improved.
Disclosure of Invention
The technical problem to be solved by the present invention is to provide an apparatus and a method for optimizing and adjusting video decoding, which can enable a video decoder to learn current decoded image content and scene classification through a neural network, and make different optimization and adjustment on classification results to achieve an optimal decoding effect.
The device of the invention is realized as follows: a video decoding optimization and adjustment device comprises a video decoding circuit, a deep learning circuit and an optimization and adjustment circuit, wherein the video decoding circuit, the deep learning circuit and the optimization and adjustment circuit are sequentially connected;
in a training mode, the video decoding circuit decodes a video source file, and sends decoded continuous video images with labels to the deep learning circuit, the deep learning circuit generates scene and content recognition results by using different convolution kernels for an original image and an inter-frame difference image, and carries out training iteration according to video labels sent by a video source and the scene and content recognition results until training is completed;
in an optimization mode, the video decoding circuit decodes a video source file, and sends decoded continuous video images to the deep learning circuit, the deep learning circuit generates scene and content identification results by using different convolution kernels for an original image and an inter-frame difference image, and sends the identification results to the optimization adjusting circuit, and the optimization adjusting circuit performs optimization adjustment processing on the video images according to the scene and content identification results and then sends the video images to the display control unit.
The method of the invention is realized as follows: a video decoding optimization adjustment method comprises a training process and an optimization process;
the training process is as follows: decoding a video source file through a video decoding circuit, sending decoded continuous video images with labels to a deep learning circuit, generating scene and content recognition results by the deep learning circuit through using different convolution kernels for an original image and an inter-frame difference image, and performing training iteration according to video labels sent by a video source and the scene and content recognition results until training is completed;
the optimization process is as follows: the video source file is decoded through a video decoding circuit, decoded continuous video images are sent to a deep learning circuit, the deep learning circuit generates scene and content identification results by using different convolution kernels for an original image and an inter-frame difference image, the identification results are sent to an optimization adjusting circuit, and the optimization adjusting circuit performs optimization adjustment processing on the video images according to the scene and content identification results and then sends the video images to a display control unit.
The invention has the following advantages: the invention can lead the neural network to learn the current decoding image content and scene classification in the video decoding process and make different optimization adjustments to the classification result so as to achieve the best decoding effect.
Drawings
The invention will be further described with reference to the following examples with reference to the accompanying drawings.
FIG. 1 is a flow chart illustrating the structure and implementation of an apparatus for video decoding optimization adjustment according to the present invention.
FIG. 2 is a block diagram of the deep learning circuit according to the present invention.
Detailed Description
Referring to fig. 1, the apparatus of the present invention is realized as follows: a video decoding optimization and adjustment device comprises a video decoding circuit, a deep learning circuit and an optimization and adjustment circuit, wherein the video decoding circuit, the deep learning circuit and the optimization and adjustment circuit are sequentially connected;
in a training mode, the video decoding circuit decodes a video source file, and sends decoded continuous video images with labels to the deep learning circuit, the deep learning circuit generates scene and content recognition results by using different convolution kernels for an original image and an inter-frame difference image, and carries out training iteration according to video labels sent by a video source and the scene and content recognition results until training is completed;
in an optimization mode, the video decoding circuit decodes a video source file, and sends decoded continuous video images to the deep learning circuit, the deep learning circuit generates scene and content identification results by using different convolution kernels for an original image and an inter-frame difference image, and sends the identification results to the optimization adjusting circuit, and the optimization adjusting circuit performs optimization adjustment processing on the video images according to the scene and content identification results and then sends the video images to the display control unit.
The invention also provides a video decoding optimization adjustment method, which comprises a training process and an optimization process;
the training process is as follows: decoding a video source file through a video decoding circuit, sending decoded continuous video images with labels to a deep learning circuit, generating scene and content recognition results by the deep learning circuit through using different convolution kernels for an original image and an inter-frame difference image, and performing training iteration according to video labels sent by a video source and the scene and content recognition results until training is completed;
the optimization process is as follows: the video source file is decoded through a video decoding circuit, decoded continuous video images are sent to a deep learning circuit, the deep learning circuit generates scene and content identification results by using different convolution kernels for an original image and an inter-frame difference image, the identification results are sent to an optimization adjusting circuit, and the optimization adjusting circuit performs optimization adjustment processing on the video images according to the scene and content identification results and then sends the video images to a display control unit.
The deep learning circuit comprises a frame buffer unit, an inter-frame difference value calculation unit, an original image neuron input unit, an inter-frame information input neuron unit and a CNN (convolutional neural network) unit; the frame buffer unit is respectively connected with the video decoding circuit, the inter-frame difference value calculation unit and the original image neuron input unit; the CNN convolutional neural network unit is respectively connected with the video decoding circuit, the original image neuron input unit, the interframe information input neuron unit and the optimization adjusting circuit.
The frame buffer unit is used for storing a current frame and a previous frame of image, sending the current frame to the optimization adjusting circuit and the original image neuron input unit, and sending the current frame and the previous frame of image to the frame difference value calculating unit;
the interframe difference value calculating unit is responsible for calculating the difference value of each pixel point position between two frames, namely the interframe difference value, and then sending the interframe difference value to the interframe information input neuron unit;
the original image neuron input unit and the interframe information input neuron unit are responsible for neuron input forming a CNN convolutional neural network;
and after receiving the input of the original image neuron input unit and the interframe information input neuron unit, the CNN convolutional neural network performs training iteration according to a video label sent by a video source and the scene and content identification result in a training mode until the training is finished.
Referring to fig. 2, the deep learning circuit further includes a parameter storage unit, where the parameter storage unit is configured to store a network weight and a threshold, an inter-frame convolution kernel, and a normal convolution kernel; the parameter storage unit may be one unit, or may be divided into three units, such as the storage network weight and threshold storage unit, the inter-frame convolution kernel storage unit, and the ordinary convolution kernel storage unit described in fig. 2.
The CNN convolutional neural network unit further comprises a neural network operation unit, a parameter initialization unit, a convolutional bias access unit, a convolutional core access unit, a weight access unit, an error calculation unit and a back propagation write-back unit; the neural network operation unit is respectively connected with the original image neuron input unit and the interframe information input neuron unit; the parameter initialization unit is connected with the parameter storage unit; the neural network operation unit is also connected with the parameter storage unit through the convolution bias access unit, the convolution kernel access unit and the weight access unit; the neural network operation unit is connected with the parameter storage unit through the error calculation unit and the back propagation write-back unit in sequence.
In the training process, the CNN convolutional neural network unit performs the following process:
(1) initialization: when training is started, the parameter initialization unit initializes three parameters of a convolution kernel, a weight value and a convolution offset value according to a preset initialization algorithm to obtain an initial value;
(2) taking the number: after initialization is completed, the original image neuron input unit and the inter-frame information input neuron unit send the original image and the inter-frame difference image to the neural network operation unit through neuron input data; meanwhile, the convolution offset access unit reads the convolution offset values of each layer of network from the parameter storage unit and sends the convolution offset values to the neural network operation unit; the convolution kernel access unit reads all convolution kernel values of each layer of network from the parameter storage unit and sends the convolution kernel values to the neural network operation unit; the weight value access unit reads the weight value of each layer of network from the parameter storage unit and sends the weight value to the neural network operation unit; when the parameter storage unit is divided into three units, namely a storage network weight and threshold storage unit, an interframe convolution kernel storage unit and a common convolution kernel storage unit, the convolution kernel access unit accesses the interframe convolution kernel storage unit and the common convolution kernel storage unit in the DDR and sends the access data to the neural network operation unit for use, wherein the interframe convolution kernel is used for an interframe difference value image input by an interframe difference value neuron, and the common convolution kernel is used for an original image input by an original image neuron;
(3) and (3) operation: after all the data fetching is finished, the neural network operation unit starts to operate according to the initial value, obtains an operation result and then sends the operation result to the error calculation unit; the error calculation unit performs error calculation according to the calculation result and the expected result, and sends the calculated error value to the back propagation write-back unit; the back propagation write-back unit calculates the update values of three parameters of a convolution kernel, a weight value and a convolution offset value according to the error value, and then writes the update values back to the corresponding parameter positions of the DDR;
and (4) after finishing one round of training, continuously repeating the steps (2) and (3) until all video source training is finished and the preset pattern recognition accuracy is reached.
As further shown in fig. 1, the optimization adjustment circuit includes a scene determination unit, a content determination unit, an image post-processing unit, and a subtitle superimposition unit; the scene judging unit and the content judging unit are both connected to the CNN convolutional neural network unit, the scene judging unit is also connected to an image post-processing unit, and the subtitle overlaying unit is connected to the content judging unit and the frame buffer unit.
The optimization adjusting circuit further comprises a scene corresponding post-processing parameter storage unit, a scene corresponding display parameter storage unit and a content corresponding subtitle storage unit, wherein the scene corresponding post-processing parameter storage unit is connected to the image post-processing unit, the scene corresponding display parameter storage unit is connected to the scene judging unit, and the content corresponding subtitle storage unit is connected to the subtitle superimposing unit.
The scene judging unit is used for sending a scene judging result generated by the convolutional neural network to the image post-processing unit;
the content judgment unit is responsible for sending a content judgment result generated by the convolutional neural network to the subtitle superposition unit;
after receiving the scene information, the image post-processing unit reads a post-processing parameter storage unit corresponding to the scene, and processes the video image, for example, if the scene is a dark night or dark image, the image contrast is integrally improved; for example, in the judgment, natural wind and light are used, the image chroma and saturation are improved, and the like; then the image after post-processing is sent to a display control unit;
the caption superposition unit is responsible for reading corresponding captions such as violent and erotic contents according to the content judgment result, and the captions can be automatically added with captions such as 'unsuitable children and please avoid children';
therefore, in the process of video decoding, the neural network can learn the current decoded image content and scene classification, and make different optimization adjustments on the classification result to achieve the optimal decoding effect.
Although specific embodiments of the invention have been described above, it will be understood by those skilled in the art that the specific embodiments described are illustrative only and are not limiting upon the scope of the invention, and that equivalent modifications and variations can be made by those skilled in the art without departing from the spirit of the invention, which is to be limited only by the appended claims.
Claims (4)
1. An apparatus for optimizing and adjusting video decoding, comprising: the video decoding circuit, the deep learning circuit and the optimization adjusting circuit are sequentially connected;
in a training mode, the video decoding circuit decodes a video source file, and sends decoded continuous video images with labels to the deep learning circuit, the deep learning circuit generates scene and content recognition results by using different convolution kernels for an original image and an inter-frame difference image, and carries out training iteration according to video labels sent by a video source and the scene and content recognition results until training is completed;
in an optimization mode, the video decoding circuit decodes a video source file and sends decoded continuous video images to the deep learning circuit, the deep learning circuit generates scene and content identification results by using different convolution kernels for an original image and an inter-frame difference image and sends the identification results to the optimization adjusting circuit, and the optimization adjusting circuit performs optimization adjustment processing on the video images according to the scene and content identification results and then sends the video images to the display control unit;
wherein the content of the first and second substances,
the deep learning circuit comprises a frame buffer unit, an inter-frame difference value calculation unit, an original image neuron input unit, an inter-frame information input neuron unit and a CNN convolution neural network unit; the frame buffer unit is respectively connected with the video decoding circuit, the inter-frame difference value calculation unit and the original image neuron input unit; the CNN convolutional neural network unit is respectively connected with the video decoding circuit, the original image neuron input unit, the interframe information input neuron unit and the optimization adjusting circuit;
the optimization adjusting circuit comprises a scene judging unit, a content judging unit, an image post-processing unit and a subtitle superposition unit; the scene judging unit and the content judging unit are both connected to the CNN convolutional neural network unit, the scene judging unit is also connected to an image post-processing unit, and the caption overlaying unit is connected to the content judging unit and the frame cache unit;
the optimization adjusting circuit further comprises a scene corresponding post-processing parameter storage unit, a scene corresponding display parameter storage unit and a content corresponding subtitle storage unit, wherein the scene corresponding post-processing parameter storage unit is connected to the image post-processing unit, the scene corresponding display parameter storage unit is connected to the scene judging unit, and the content corresponding subtitle storage unit is connected to the subtitle superimposing unit;
the scene judging unit sends a scene judging result generated by the convolutional neural network to the image post-processing unit; after receiving the scene information, the image post-processing unit reads a post-processing parameter storage unit corresponding to the scene, processes the video image, and integrally improves the image contrast if the scene is a dark night or dark image; if the scene is natural scene, the image chroma and saturation are improved, and then the image chroma and saturation are sent to a display control unit; the content judgment unit sends a content judgment result generated by the convolutional neural network to the caption superposition unit; the caption superposition unit reads the corresponding caption according to the content judgment result, and if the content judgment result is violent and erotic content, the caption automatically adds the caption of 'unsuitable children and please avoid children'.
2. The video decoding optimization adjustment apparatus according to claim 1, wherein: the deep learning circuit also comprises a parameter storage unit, wherein the parameter storage unit is used for storing network weight and threshold, interframe convolution kernel and common convolution kernel;
the CNN convolutional neural network unit further comprises a neural network operation unit, a parameter initialization unit, a convolutional bias access unit, a convolutional core access unit, a weight access unit, an error calculation unit and a back propagation write-back unit; the neural network operation unit is respectively connected with the original image neuron input unit and the interframe information input neuron unit; the parameter initialization unit is connected with the parameter storage unit; the neural network operation unit is also connected with the parameter storage unit through the convolution bias access unit, the convolution kernel access unit and the weight access unit; the neural network operation unit is connected with the parameter storage unit through the error calculation unit and the back propagation write-back unit in sequence.
3. A video decoding optimization adjustment method is characterized in that: the method comprises a training process and an optimization process;
the training process is as follows: decoding a video source file through a video decoding circuit, sending decoded continuous video images with labels to a deep learning circuit, generating scene and content recognition results by the deep learning circuit through using different convolution kernels for an original image and an inter-frame difference image, and performing training iteration according to video labels sent by a video source and the scene and content recognition results until training is completed;
the optimization process is as follows: decoding a video source file through a video decoding circuit, and sending decoded continuous video images to a deep learning circuit, wherein the deep learning circuit generates scene and content identification results by using different convolution kernels for an original image and an inter-frame difference image, and sends the identification results to an optimization adjusting circuit, and the optimization adjusting circuit performs optimization adjustment processing on the video images according to the scene and content identification results and then sends the video images to a display control unit;
the deep learning circuit comprises a frame buffer unit, an inter-frame difference value calculation unit, an original image neuron input unit, an inter-frame information input neuron unit and a CNN convolution neural network unit; the frame buffer unit is used for storing a current frame and a previous frame of image, sending the current frame to the optimization adjusting circuit and the original image neuron input unit, and sending the current frame and the previous frame of image to the frame difference value calculating unit;
the interframe difference value calculating unit is responsible for calculating the difference value of each pixel point position between two frames, namely the interframe difference value, and then sending the interframe difference value to the interframe information input neuron unit; the original image neuron input unit and the interframe information input neuron unit are responsible for neuron input forming a CNN convolutional neural network; after receiving the input of the original image neuron input unit and the interframe information input neuron unit, the CNN convolutional neural network performs training iteration according to a video label sent by a video source and the scene and content identification result in a training mode until the training is finished;
the optimization adjusting circuit comprises a scene judging unit, a content judging unit, an image post-processing unit and a subtitle superposition unit; the scene judging unit and the content judging unit are both connected to the CNN convolutional neural network unit, the scene judging unit is also connected to an image post-processing unit, and the caption overlaying unit is connected to the content judging unit and the frame cache unit;
the optimization adjusting circuit further comprises a scene corresponding post-processing parameter storage unit, a scene corresponding display parameter storage unit and a content corresponding subtitle storage unit, wherein the scene corresponding post-processing parameter storage unit is connected to the image post-processing unit, the scene corresponding display parameter storage unit is connected to the scene judging unit, and the content corresponding subtitle storage unit is connected to the subtitle superimposing unit;
the scene judging unit sends a scene judging result generated by the convolutional neural network to the image post-processing unit; after receiving the scene information, the image post-processing unit reads a post-processing parameter storage unit corresponding to the scene, processes the video image, and integrally improves the image contrast if the scene is a dark night or dark image; if the scene is natural scene, the image chroma and saturation are improved, and then the image chroma and saturation are sent to a display control unit; the content judgment unit sends a content judgment result generated by the convolutional neural network to the caption superposition unit; the caption superposition unit reads the corresponding caption according to the content judgment result, and if the content judgment result is violent and erotic content, the caption automatically adds the caption of 'unsuitable children and please avoid children'.
4. The video decoding optimization adjustment method according to claim 3, wherein: the deep learning circuit also comprises a parameter storage unit and a scene corresponding post-processing parameter storage unit, wherein the parameter storage unit is used for storing network weight and threshold, interframe convolution kernel and common convolution kernel; the CNN convolutional neural network unit further comprises a neural network operation unit, a parameter initialization unit, a convolutional bias access unit, a convolutional core access unit, a weight access unit, an error calculation unit and a back propagation write-back unit; in the training process, the CNN convolutional neural network unit performs the following process:
(1) initialization: when training is started, the parameter initialization unit initializes three parameters of a convolution kernel, a weight value and a convolution offset value according to a preset initialization algorithm to obtain an initial value;
(2) taking the number: after initialization is completed, the original image neuron input unit and the inter-frame information input neuron unit send the original image and the inter-frame difference image to the neural network operation unit through neuron input data; meanwhile, the convolution offset access unit reads the convolution offset values of each layer of network from the parameter storage unit and sends the convolution offset values to the neural network operation unit; the convolution kernel access unit reads all convolution kernel values of each layer of network from the parameter storage unit and sends the convolution kernel values to the neural network operation unit; the weight value access unit reads the weight value of each layer of network from the parameter storage unit and sends the weight value to the neural network operation unit;
(3) and (3) operation: after all the data fetching is finished, the neural network operation unit starts to operate according to the initial value, obtains an operation result and then sends the operation result to the error calculation unit; the error calculation unit performs error calculation according to the calculation result and the expected result, and sends the calculated error value to the back propagation write-back unit; the back propagation write-back unit calculates the update values of three parameters of a convolution kernel, a weight value and a convolution offset value according to the error value, and then writes the update values back to the corresponding parameter positions of the DDR;
and (4) after finishing one round of training, continuously repeating the steps (2) and (3) until all video source training is finished and the preset pattern recognition accuracy is reached.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810051581.XA CN108462876B (en) | 2018-01-19 | 2018-01-19 | Video decoding optimization adjustment device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810051581.XA CN108462876B (en) | 2018-01-19 | 2018-01-19 | Video decoding optimization adjustment device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108462876A CN108462876A (en) | 2018-08-28 |
| CN108462876B true CN108462876B (en) | 2021-01-26 |
Family
ID=63220994
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810051581.XA Active CN108462876B (en) | 2018-01-19 | 2018-01-19 | Video decoding optimization adjustment device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108462876B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109840590A (en) * | 2019-01-31 | 2019-06-04 | 福州瑞芯微电子股份有限公司 | A kind of scene classification circuit framework neural network based and method |
| CN110458051A (en) * | 2019-07-25 | 2019-11-15 | 中移（杭州）信息技术有限公司 | A kind of method, apparatus, server and the readable storage medium storing program for executing of equipment control |
| CN112631120B (en) * | 2019-10-09 | 2022-05-17 | Oppo广东移动通信有限公司 | PID control method, device and video coding and decoding system |
| CN110493639A (en) * | 2019-10-21 | 2019-11-22 | 南京创维信息技术研究院有限公司 | A kind of method and system of adjust automatically sound and image model based on scene Recognition |
| CN112929703A (en) * | 2019-12-06 | 2021-06-08 | 上海海思技术有限公司 | Method and device for processing code stream data |
| CN111031392A (en) * | 2019-12-23 | 2020-04-17 | 广州视源电子科技股份有限公司 | Media file playing method, system, device, storage medium and processor |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102158699A (en) * | 2011-03-28 | 2011-08-17 | 广州市聚晖电子科技有限公司 | Embedded video compression coding system with image enhancement function |
| CN104299196A (en) * | 2014-10-11 | 2015-01-21 | 京东方科技集团股份有限公司 | Image processing device and method and display device |
| CN104881675A (en) * | 2015-05-04 | 2015-09-02 | 北京奇艺世纪科技有限公司 | Video scene identification method and apparatus |
| CN105302872A (en) * | 2015-09-30 | 2016-02-03 | 努比亚技术有限公司 | Image processing device and method |
| CN107590443A (en) * | 2017-08-23 | 2018-01-16 | 上海交通大学 | Limiter stage live video automatic testing method and system based on the study of depth residual error |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014295B (en) * | 2010-11-19 | 2012-11-28 | 嘉兴学院 | Network sensitive video detection method |
| CN102254224A (en) * | 2011-07-06 | 2011-11-23 | 无锡泛太科技有限公司 | Internet of things electric automobile charging station system based on image identification of rough set neural network |
| US9143823B2 (en) * | 2012-10-01 | 2015-09-22 | Google Inc. | Providing suggestions for optimizing videos to video owners |
| US10521671B2 (en) * | 2014-02-28 | 2019-12-31 | Second Spectrum, Inc. | Methods and systems of spatiotemporal pattern recognition for video content development |
| CN104202604B (en) * | 2014-08-14 | 2017-09-22 | 深圳市腾讯计算机系统有限公司 | The method and apparatus of video source modeling |
| CN104933680B (en) * | 2015-03-13 | 2017-10-31 | 哈尔滨工程大学 | A kind of intelligent quick sea fog minimizing technology of unmanned boat vision system video |
| US20170148487A1 (en) * | 2015-11-25 | 2017-05-25 | Dell Products L.P. | Video Manipulation for Privacy Enhancement |
-
2018
- 2018-01-19 CN CN201810051581.XA patent/CN108462876B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102158699A (en) * | 2011-03-28 | 2011-08-17 | 广州市聚晖电子科技有限公司 | Embedded video compression coding system with image enhancement function |
| CN104299196A (en) * | 2014-10-11 | 2015-01-21 | 京东方科技集团股份有限公司 | Image processing device and method and display device |
| CN104881675A (en) * | 2015-05-04 | 2015-09-02 | 北京奇艺世纪科技有限公司 | Video scene identification method and apparatus |
| CN105302872A (en) * | 2015-09-30 | 2016-02-03 | 努比亚技术有限公司 | Image processing device and method |
| CN107590443A (en) * | 2017-08-23 | 2018-01-16 | 上海交通大学 | Limiter stage live video automatic testing method and system based on the study of depth residual error |
Non-Patent Citations (1)
| Title |
|---|
| Video pornography detection through deep learning techniques and motion;Mauricio Perez等;《Neurocomputing》;20170322(第230期);第279-291页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108462876A (en) | 2018-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108462876B (en) | Video decoding optimization adjustment device and method | |
| TWI826321B (en) | A method for enhancing quality of media | |
| CN110324664B (en) | Video frame supplementing method based on neural network and training method of model thereof | |
| CN110046546B (en) | Adaptive sight tracking method, device and system and storage medium | |
| US20190180454A1 (en) | Detecting motion dragging artifacts for dynamic adjustment of frame rate conversion settings | |
| EP3707906B1 (en) | Electronic apparatus and control method thereof | |
| US11490157B2 (en) | Method for controlling video enhancement, device, electronic device and storage medium | |
| CN109698957B (en) | Image coding method and device, computing equipment and storage medium | |
| WO2021098616A1 (en) | Motion posture recognition method, motion posture recognition apparatus, terminal device and medium | |
| CN111047543A (en) | Image enhancement method, device and storage medium | |
| US20200177470A1 (en) | Method for enhancing quality of media | |
| CN109035147A (en) | Image processing method and device, electronic device, storage medium and computer equipment | |
| CN110620924A (en) | Method and device for processing coded data, computer equipment and storage medium | |
| CN111401259A (en) | Model training method, system, computer readable medium and electronic device | |
| US20220398692A1 (en) | Video conferencing based on adaptive face re-enactment and face restoration | |
| CN116205820A (en) | Image enhancement method, target identification method, device and medium | |
| CN110503002B (en) | Face detection method and storage medium | |
| CN108961314B (en) | Moving image generation method, moving image generation device, electronic device, and computer-readable storage medium | |
| US20220044423A1 (en) | Ball trajectory tracking | |
| WO2018057449A1 (en) | Auto-directing media construction | |
| CN113162895B (en) | Dynamic coding method, streaming media quality determination method and electronic equipment | |
| CN107622248B (en) | Gaze identification and interaction method and device | |
| CN110971924B (en) | Method, device, storage medium and system for beautifying in live broadcast process | |
| CN111414959A (en) | Image recognition method and device, computer readable medium and electronic equipment | |
| CN113891069A (en) | Video quality assessment method, device and equipment |
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 | ||
| CB02 | Change of applicant information |
Address after: 350000 building 18, 89 software Avenue, Gulou District, Fuzhou City, Fujian Province Applicant after: Ruixin Microelectronics Co., Ltd Address before: 350000 building 18, 89 software Avenue, Gulou District, Fuzhou City, Fujian Province Applicant before: Fuzhou Rockchips Electronics Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |