CN109862207B - KVM video content change detection method based on compressed domain - Google Patents
KVM video content change detection method based on compressed domain Download PDFInfo
- Publication number
- CN109862207B CN109862207B CN201910107774.7A CN201910107774A CN109862207B CN 109862207 B CN109862207 B CN 109862207B CN 201910107774 A CN201910107774 A CN 201910107774A CN 109862207 B CN109862207 B CN 109862207B
- Authority
- CN
- China
- Prior art keywords
- frame
- current frame
- content change
- macro blocks
- code stream
- 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
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention relates to a KVM video content change detection method based on a compressed domain, which comprises the following steps: (1) inputting a KVM video code stream of a group of images; (2) analyzing the video code stream to obtain compressed domain coding information of a P frame; (3) filtering the bit number of the macro block; (4) judging whether the content of the current frame changes or not in multiple stages by utilizing the coding information; (5) the group of images is saved or discarded according to the content change flag. The method fully utilizes the coding information generated in the video coding process, does not need to reconstruct images, can effectively detect the change of the KVM video content by taking the image group as a unit, greatly reduces the calculation complexity compared with a complete pixel domain processing method, and simultaneously effectively reduces the storage quantity of the KVM video.
Description
Technical Field
The invention relates to the field of digital video processing, in particular to a KVM video content change detection method based on a compressed domain.
Background
A Keyboard (Keyboard), display (Video) and Mouse (Mouse) remote switch management system, i.e., a KVM system, also called a multi-computer control system, is mainly used for remote control of various computers in a computer room. The KVM video is data acquired and compressed from an original video output by a computer screen in a digital KVM system, and in recent years, with the development of high-performance computers and high-bandwidth networks, the application of the KVM system is continuously increased, and the storage and management of the large amount of stored KVM videos are greatly inconvenient. In order to reduce the storage capacity of the KVM video and facilitate management, it is important to detect the change of the KVM video content. A plurality of redundant parts without content change exist in the KVM video, and content change detection is carried out on the redundant parts, so that only the parts with the changed video content can be stored, and the redundant parts without the changed video content are discarded, thereby reducing the storage capacity and facilitating the data management in the later period. In addition, KVM video content change detection is widely applied in some specific fields, for example, some industrial control devices may pop up an alarm popup window when working, and the alarm popup window may be detected by using a content change detection method and processed in time, so that some unnecessary losses may be reduced.
At present, most of the huge KVM videos are encoded and transmitted by the mainstream H.264 standard. In order to reduce the delay of the KVM system, in h.264, mainly an I frame and a P frame are used for encoding transmission, the I frame is also called an intra-frame encoded frame, which is an independent frame with all information, and can be independently decoded without referring to other images, and the P frame is also called an inter-frame predictive encoded frame, which needs to refer to a forward reconstructed frame for encoding. The h.264 compression standard uses 16 × 16 macroblocks, and the coding types of the macroblocks are classified into Intra macroblocks and Inter macroblocks. The Intra macroblock is predicted from the reconstructed data in the current frame, and the Inter macroblock is predicted using the reconstructed reference frame image, and its motion vector refers to the offset of the current macroblock from the reference frame image, and includes a horizontal component and a vertical component.
At present, the existing video content change detection technologies mainly include background subtraction, frame difference, background modeling, feature-based motion detection, and optical flow-based motion detection. Patent with application number CN201210334876 discloses a method for detecting change of video images output by a computer. The method comprises the steps of firstly converting a color image into a gray image; then mean value filtering is carried out on the gray level image; then, carrying out blocking, frame difference, mean value and variance on the filtered image; and finally, calculating a discrimination threshold value according to the mean value and the variance so as to judge whether the variation occurs. Patent No. CN201110009716 discloses an image change detection method based on principal component generalized inverse transformation. The method comprises the steps of carrying out transformation in two feature spaces after reorganizing data of an image to be detected, carrying out updating wave band difference processing in the transformed feature spaces to obtain change components, and extracting a change area through an automatic threshold value method to realize image change detection. Patent application No. CN200810101240.5 discloses a method for detecting moving objects in video according to scene change characteristics. Firstly, carrying out feature point detection, background feature point parameters and model pixel point parameter calculation according to a training image, and calculating motion feature points by adopting the background feature point parameters and the currently detected image feature points; classifying pixel point parameters in the background model by using the motion characteristic points; calculating the updating rate of the background model pixel point parameters by using the classification result; and finally, obtaining the position and the shape of the moving object by adopting a background difference method according to the current image and the background model pixel point parameters.
The method mainly comprises the steps of carrying out video content change detection in a pixel domain, so that the overall calculation complexity is high, and in addition, for the KVM video, the decoding reconstruction of the pixel domain is required, so that the calculation amount is further increased, and therefore, the method cannot be well suitable for the content change detection of the KVM video. In the H.264 code stream, the code information such as the frame bit number, the macro block type, the motion vector and the like which reflect the content change is contained, and the code stream can be only analyzed without complete reconstruction, so that the content change detection of the KVM video in the H.264 compressed domain can ensure lower calculation complexity.
Disclosure of Invention
In order to reduce the data amount stored in the KVM video code stream, the invention provides a method for detecting the change of the KVM video content based on a compression domain by analyzing and obtaining coding information from the code stream, which comprises the following steps:
(1) KVM video code stream for inputting a group of images
The input KVM video code stream is H.264 code stream, the supported frame types are I frame (intra-frame coding frame) and P frame (forward inter-frame coding frame), one group of pictures (GOP) is composed of one I frame and a plurality of subsequent P frames, the P frames do not perform forced macroblock refreshing, namely, an encoder does not forcibly appoint the coding type of the macroblock, but are preferably obtained from various coding types, so that the coding type of the macroblock of the P frame can reflect the difference degree of the image content of the macroblock of the P frame and the previous frame.
(2) Analyzing video code stream to obtain compressed domain coding information of P frame
The P frame coding information obtained by analysis comprises coding types, bit numbers and motion vectors of all macro blocks of the frame; if the frame is the first P frame, setting a content change mark CF as 0, and analyzing a P frame as the current frame; the coding types of the macro blocks comprise Intra macro blocks (Intra-frame coding macro blocks) and Inter macro blocks (Inter-frame coding macro blocks), according to statistics, the Intra macro blocks in the P frame mostly appear in areas with obviously changed video contents, and the Inter macro blocks are mostly used in other cases. The I frame has no Inter-frame prediction coding information, only has an Intra macro block and no Inter macro block, and cannot predict the image content difference degree of the previous frame and the next frame according to the macro block type, so the invention does not process the I frame.
(3) Macroblock bit number filtering
At present, most of KVM video code streams are obtained by coding after VGA interface module conversion, and the coded video has some noises which influence the detection of content change, so that the coding information needs to be filtered. The bit number of the completely static macro block is equal to or close to zero through analyzing the code stream, the bit number of the macro block with obvious content change is larger, and the bit number of the static macro block with partial noise is nonzero and smaller. According to these features, the invention filters the number of macroblock bits: will be less than the threshold THbit1The bit number of the macro block is directly set to be zero, and TH is obtained by analyzing and counting the bit number of the macro block with noise in the KVM video code streambit1Has a value range of [50,300 ]]And the user selects according to the noise condition of the actual code stream, and the noise is large in value.
(4) Multi-level judgment on whether the content of the current frame changes or not by utilizing coding information
The method comprises the following specific steps of judging whether video content changes or not by using coding information such as macro block coding types, macro block bit numbers, motion vectors and the like of a current frame and a previous frame aiming at a KVM typical scene:
(4-1) determination of whether there is significant change in the entire desktop
If the number of the Intra macro blocks of the current frame is larger than that of the Intra macro blocks of the previous frame, and the percentage of the number of the Intra macro blocks of the current frame to the total number of the macro blocks is larger than a threshold THintra1Then, the current frame is describedSetting a content change mark CF as 1 when the content is greatly changed compared with the previous frame, and jumping to the step (5), otherwise, carrying out the next detection; obtaining TH by statistically analyzing the number of the Intra macro blocks of the P frame in the scene with significant changes on the whole desktopintra1Has a value range of [40,90 ]]。
(4-2) determination of whether operation frame change occurs on computer desktop
If the number of the Intra macro blocks of the current frame is larger than that of the Intra macro blocks of the previous frame and the difference value is larger than the threshold value THintra2And the maximum number of Intra macro blocks with connected four adjacent domains in the current frame is greater than the threshold value THintra3If yes, the image content of a certain area in the current frame is obviously changed, a content change mark CF is set to be 1, the step (5) is carried out, and if not, the next detection is carried out; TH is obtained by statistically analyzing the number of Intra macro blocks in a P frame of an operation frame appearing on a computer desktop, the number of Intra macro blocks in a previous frame of the operation frame, and the number of maximum Intra macro blocks communicated with four adjacent domains in the frameintra2Has a value range of [20,1000%]、THintra3Has a value range of [3,300 ]]And selecting according to the actual image resolution to be detected and the size of the operation frame, wherein the larger the image resolution is, the larger the target operation frame is, and the larger the value is. The four-adjacent domain communication is that the current macro block is communicated with the upper macro block, the lower macro block, the left macro block or the right macro block, and is a classical communication mode in the field of image processing.
(4-3) discrimination of Small changes in mouse and characters
If the number of bits existing in the current frame is larger than the threshold THbit2The number of the macro blocks is larger than a threshold THmb1Or the bit number of at least 1 Intra macro block in the current frame is larger than the threshold value THbit2If so, the two conditions indicate that the image content of the current frame is significantly changed in a tiny area compared with the previous frame, the bit number of the corresponding macro block is particularly large, so that the content change flag CF is set to be 1, the step (5) is skipped, and if not, the next detection is performed; TH is obtained by statistically analyzing the bit number of macro blocks in a P frame with mouse movement and character input on a computer desktop and adjacent macro blocks in four adjacent domains of the P framebit2Has a value range of [200, 1000%]、THmb1Has a value range of [1,4 ]]。
(4-4) determination of whether or not there is a change such as a global movement in the application software
Firstly, the motion vector amplitude MV of each macro block of the current frame is calculated according to the formula (1)apWherein | MVx| and | MVyI is the horizontal component amplitude and the vertical component amplitude of the motion vector of each macro block of the current frame obtained by analysis in the step (2); if MV exists in the current frameapAre all greater than a threshold value THmvThe number of the macro blocks is larger than a threshold THmb2If so, the change of the whole movement of the content transmission of the partial area of the current frame is described, the amplitude of the motion vector is larger, and a content change mark CF is set to be 1; TH is obtained by statistically analyzing the macro block motion vector with larger amplitude in the P frame with the change of the whole movement of the application software and the like and the adjacent macro blocks in four adjacent domains thereofmvHas a value range of [16,128 ]]、THmb2Has a value range of [4,100 ]](ii) a Intra macroblock motion vector free, MVapAre set to 0.
MVap＝|MVx|+|MVy| (1)
(5) Saving or discarding group of pictures code stream according to content change flag
If the content change flag CF of the current frame is 1, which indicates that the content of at least one frame in the current image group changes, the code stream of the subsequent frame is not analyzed and judged any more, and the video code stream of the current image group is stored; if the content change flag CF of the current frame is 0 and the current frame is the last P frame of the current image group, discarding the video code stream of the current image group; and (4) if the content change flag CF of the current frame is 0 and the current frame is not the last P frame of the current image group, repeating the steps (2) to (4) to process the detection of the next P frame.
Compared with the prior art, the invention has the following beneficial effects:
the compressed domain coding information contained in the H.264 code stream is directly utilized to detect the change of the KVM video content, the image reconstruction is not needed, the content change can be rapidly detected by taking the image group as a unit, the storage data volume of the KVM video is effectively reduced, and the calculation complexity is greatly reduced compared with a pixel domain detection method.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Detailed Description
The present invention will be described in detail below with reference to examples and drawings, but the present invention is not limited thereto. The video decoder adopted in the embodiment is FFmpeg, and the version number is 3.2.4; the input video source is a video recorded by a KVM switcher from a VGA port, the resolution ratio is 1080P, the recorded target code rate is 4Mbps, the frame rate is 30fps, the coding prediction structure is an IPPP mode, and the I-frame interval is 150; the recorded KVM video content comprises scenes such as mouse movement, popup, Chinese character input, word document sliding, rapid and large desktop change and the like. As shown in fig. 1, a KVM video content change detection method based on compressed domain includes the following steps:
(1) inputting a KVM video code stream of a group of images;
(2) analyzing the video code stream to obtain compressed domain coding information of a P frame;
(3) filtering the bit number of the macro block;
(4) judging whether the content of the current frame changes or not in multiple stages by utilizing the coding information;
(5) the group of images is saved or discarded according to the content change flag.
The step (1) specifically comprises the following steps:
inputting a video code stream of a group of pictures, wherein the video code stream of the group of pictures is an H.264 code stream, the first one is an I frame and the subsequent 149P frames.
The step (2) specifically comprises the following steps:
the P frame coding information obtained by analysis comprises coding types, bit numbers and motion vectors of all macro blocks of the frame; if the frame is the first P frame, setting a content change mark CF as 0, and analyzing a P frame as the current frame; the coding types of the macro blocks include Intra macro blocks and Inter macro blocks. The present invention does not process I frames.
The step (3) specifically comprises the following steps:
the bit number occupied by each macro block in the current frame is counted, and the counted number is smaller than a threshold THbit1The number of bits of the macro block is directly set toZero; wherein TH isbit1Has a value range of [50,300 ]]Here, 200 is taken.
The step (4) specifically comprises the following steps:
the method for judging whether the video content changes or not by using the macro block coding type, the macro block bit number, the motion vector and other coding information of the current frame and the previous frame aiming at the KVM typical scene specifically comprises the following steps:
(4-1) determination of whether there is significant change in the entire desktop
If the number of the Intra macro blocks of the current frame is larger than that of the Intra macro blocks of the previous frame, and the percentage of the number of the Intra macro blocks of the current frame to the total number of the macro blocks is larger than a threshold THintra1If yes, setting the content change mark CF as 1, and jumping to the step (5), otherwise, carrying out the next detection; wherein TH isintra1Has a value range of [40,90 ]]Here 70 is taken.
(4-2) determination of whether operation frame change occurs on computer desktop
If the number of the Intra macro blocks of the current frame is larger than that of the Intra macro blocks of the previous frame and the difference value is larger than the threshold value THintra2And the maximum number of Intra macro blocks with connected four adjacent domains in the current frame is greater than the threshold value THintra3If yes, setting the content change mark CF as 1, and jumping to the step (5), otherwise, carrying out the next detection; wherein TH isintra2Has a value range of [20,1000%]Where 80, TH are takenintra3Has a value range of [3,300 ]]Here 20 is taken.
(4-3) discrimination of Small changes in mouse and characters
If the number of bits existing in the current frame is larger than the threshold THbit2The number of the macro blocks is larger than a threshold THmb1Or the bit number of at least 1 Intra macro block in the current frame is larger than the threshold value THbit2If so, setting the content change mark CF to be 1 under both the two conditions, and jumping to the step (5), otherwise, carrying out the next detection; wherein TH isbit2Has a value range of [200, 1000%]Here, take 300, THmb1Has a value range of [1,4 ]]Here, 2 is taken.
(4-4) determination of whether or not there is a change such as a global movement in the application software
Firstly, according to formula (1) to calculateMotion vector magnitude MVa to each macroblock of the current framepWherein | MVx| and | MVyI is the horizontal component amplitude and the vertical component amplitude of the motion vector of each macro block of the current frame obtained by analysis in the step (2); if MV exists in the current frameapAre all greater than a threshold value THmvThe number of the macro blocks is larger than a threshold THmb2If yes, setting the content change flag CF as 1; wherein TH ismvHas a value range of [16,128 ]]Here, take 50, THmb2Has a value range of [4,100 ]]Here, 10 is taken; intra macroblock motion vector free, MVapAre set to 0.
MVap＝|MVx|+|MVy| (1)
The step (5) specifically comprises the following steps:
if the content change flag CF of the current frame is 1, which indicates that the content of at least one frame in the current image group changes, the code stream of the subsequent frame is not analyzed and judged any more, and the video code stream of the current image group is stored; if the content change flag CF of the current frame is 0 and the current frame is the last P frame of the current image group, discarding the video code stream of the current image group; and (4) if the content change flag CF of the current frame is 0 and the current frame is not the last P frame of the current image group, repeating the steps (2) to (4) to process the detection of the next P frame.
Claims (2)
1. A KVM video content change detection method based on compressed domain is characterized in that the detection method comprises the following steps:
(1) inputting a KVM video code stream of a group of images:
the input KVM video code stream is H.264 code stream, one image group consists of one I frame and a plurality of subsequent P frames, and the P frames are not subjected to forced macro block refreshing;
(2) analyzing the video code stream to obtain the compressed domain coding information of a P frame:
the compressed domain coding information of the P frame obtained by analysis comprises the coding types, the bit numbers and the motion vectors of all macro blocks of the frame; if the frame is the first P frame, setting a content change mark CF as 0, and analyzing a P frame as the current frame; the coding types of the macro blocks comprise Intra-frame coding macro block Intra macro blocks and Inter-frame coding macro block Inter macro blocks;
(3) and (3) macroblock bit number filtering:
the bit number occupied by each macro block in the current frame is counted, and the counted number is smaller than a threshold THbit1The bit number of the macro block is directly set to zero;
(4) judging whether the content of the current frame changes or not in multiple stages by utilizing the coding information:
judging whether the KVM video content has change or not by using the macro block coding type, the bit number and the motion vector of the current frame and the previous frame; the step (4) comprises the following substeps:
(4-1) if the number of the Intra macro blocks of the current frame is larger than that of the Intra macro blocks of the previous frame, and the percentage of the number of the Intra macro blocks of the current frame to the total number of the macro blocks is larger than a threshold THintra1If yes, setting the content change mark CF as 1, jumping to the step (5), otherwise, carrying out the next detection; wherein TH isintra1Has a value range of [40,90 ]]；
(4-2) if the number of the Intra macro blocks of the current frame is larger than that of the Intra macro blocks of the previous frame and the difference value is larger than the threshold value THintra2And the maximum number of Intra macro blocks with connected four adjacent domains in the current frame is greater than the threshold value THintra3If yes, setting the content change mark CF as 1, and jumping to the step (5), otherwise, carrying out the next detection; wherein TH isintra2Has a value range of [20,1000%]、THintra3Has a value range of [3,300 ]]；
(4-3) if the bit numbers of the current frame are all larger than the threshold value THbit2The number of the macro blocks is larger than a threshold THmb1Or the bit number of at least 1 Intra macro block in the current frame is larger than the threshold value THbit2If yes, setting the content change mark CF as 1, and jumping to the step (5), otherwise, carrying out the next detection; wherein TH ismb1Has a value range of [1,4 ]]、THbit2Has a value range of [200, 1000%]；
(4-4) firstly, calculating the motion vector amplitude MV of each macro block of the current frame according to the formula (1)apWherein | MVx| and | MVyThe | is the motion vector of each macro block of the current frame obtained by analysis in the step (2)A horizontal component amplitude and a vertical component amplitude,
MVap＝|MVx|+|MVy| (1)
if MV exists in the current frameapAre all greater than a threshold value THmvThe number of the macro blocks is larger than a threshold THmb2If yes, setting the content change flag CF as 1; wherein TH ismvHas a value range of [16,128 ]]、THmb2Has a value range of [4,100 ]]；
(5) Saving or discarding the group of images according to the content change flag:
if the content change mark CF of the current frame is 1, no further code stream analysis and judgment of the subsequent frame is carried out, and the video code stream of the current image group is stored; if the content change flag CF of the current frame is 0 and the current frame is the last P frame of the current image group, discarding the video code stream of the current image group; and (4) if the content change flag CF of the current frame is 0 and the current frame is not the last P frame of the current image group, repeating the steps (2) to (4) to process the detection of the next P frame.
2. The KVM video content change detecting method based on compressed domain as claimed in claim 1, wherein in step (3), THbit1Has a value range of [50,300 ]]。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910107774.7A CN109862207B (en) | 2019-02-02 | 2019-02-02 | KVM video content change detection method based on compressed domain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910107774.7A CN109862207B (en) | 2019-02-02 | 2019-02-02 | KVM video content change detection method based on compressed domain |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109862207A CN109862207A (en) | 2019-06-07 |
| CN109862207B true CN109862207B (en) | 2021-04-09 |
Family
ID=66897566
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910107774.7A Active CN109862207B (en) | 2019-02-02 | 2019-02-02 | KVM video content change detection method based on compressed domain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109862207B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112637538B (en) * | 2020-12-15 | 2022-08-02 | 重庆紫光华山智安科技有限公司 | Smart tag method, system, medium, and terminal for optimizing video analysis |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101090498A (en) * | 2007-07-19 | 2007-12-19 | 华为技术有限公司 | Device and method for motion detection of image |
| CN101246547A (en) * | 2008-03-03 | 2008-08-20 | 北京航空航天大学 | Method for detecting moving objects in video according to scene variation characteristic |
| CN102063722A (en) * | 2011-01-18 | 2011-05-18 | 上海交通大学 | Image change detecting method based on principle component general inverse transformation |
| CN102867308A (en) * | 2012-09-11 | 2013-01-09 | 大连理工大学 | Method for detecting change of video image output by computer |
| CN103985250A (en) * | 2014-04-04 | 2014-08-13 | 浙江工业大学 | Light-weight holographic road traffic state visual inspection device |
| CN104683802A (en) * | 2015-03-24 | 2015-06-03 | 江南大学 | H.264/AVC compressed domain based moving target tracking method |
| CN109104609A (en) * | 2018-09-12 | 2018-12-28 | 浙江工业大学 | A kind of lens boundary detection method merging HEVC compression domain and pixel domain |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060109902A1 (en) * | 2004-11-19 | 2006-05-25 | Nokia Corporation | Compressed domain temporal segmentation of video sequences |
-
2019
- 2019-02-02 CN CN201910107774.7A patent/CN109862207B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101090498A (en) * | 2007-07-19 | 2007-12-19 | 华为技术有限公司 | Device and method for motion detection of image |
| CN101246547A (en) * | 2008-03-03 | 2008-08-20 | 北京航空航天大学 | Method for detecting moving objects in video according to scene variation characteristic |
| CN102063722A (en) * | 2011-01-18 | 2011-05-18 | 上海交通大学 | Image change detecting method based on principle component general inverse transformation |
| CN102867308A (en) * | 2012-09-11 | 2013-01-09 | 大连理工大学 | Method for detecting change of video image output by computer |
| CN103985250A (en) * | 2014-04-04 | 2014-08-13 | 浙江工业大学 | Light-weight holographic road traffic state visual inspection device |
| CN104683802A (en) * | 2015-03-24 | 2015-06-03 | 江南大学 | H.264/AVC compressed domain based moving target tracking method |
| CN109104609A (en) * | 2018-09-12 | 2018-12-28 | 浙江工业大学 | A kind of lens boundary detection method merging HEVC compression domain and pixel domain |
Non-Patent Citations (1)
| Title |
|---|
| 《一种HEVC压缩域的运动目标检测方法》;杨洋 等;《小型微型计算机系统》;20180711;第39卷(第5期);1079-1084 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109862207A (en) | 2019-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110300977B (en) | Method for image processing and video compression | |
| EP2782340B1 (en) | Motion analysis method based on video compression code stream, code stream conversion method and apparatus thereof | |
| CN108495135B (en) | Quick coding method for screen content video coding | |
| CN101583036B (en) | Method for determining the relation between movement characteristics and high efficient coding mode in pixel-domain video transcoding | |
| WO2023016155A1 (en) | Image processing method and apparatus, medium, and electronic device | |
| KR101149522B1 (en) | Apparatus and method for detecting scene change | |
| US9712828B2 (en) | Foreground motion detection in compressed video data | |
| CN109862207B (en) | KVM video content change detection method based on compressed domain | |
| CN111770334B (en) | Data encoding method and device, and data decoding method and device | |
| CN101425129B (en) | Target abnormal detecting method and device based on JPEG image | |
| JP7411785B2 (en) | Interpolation filtering method and device, computer program and electronic device for intra prediction | |
| CN105359508A (en) | Multi-level spatial-temporal resolution increase of video | |
| CN104410863B (en) | Image processor and image processing method | |
| CN114793282A (en) | Neural network based video compression with bit allocation | |
| CN107087173B (en) | A kind of video encoding optimization method of content oriented analysis | |
| CN1338235A (en) | Method and device for video compress encoding based on division | |
| US11622118B2 (en) | Determination of coding modes for video content using order of potential coding modes and block classification | |
| CN116437102B (en) | Method, system, equipment and storage medium for learning universal video coding | |
| US20240121395A1 (en) | Methods and non-transitory computer readable storage medium for pre-analysis based resampling compression for machine vision | |
| CN107277508B (en) | Pixel-level bidirectional intra-frame prediction method adopting self-adaptive mode selection | |
| Makarov et al. | Binary shape coding using finite automata | |
| Lin et al. | Hierarchical Bi-directional Temporal Context Mining for Improved Video Compression | |
| Kim et al. | Moving Object Detection Using Syntax Elements | |
| CN117750006A (en) | HEVC video double-compression detection method under different bit rates | |
| Xiao-li | Research and Implementation of the Video I-frame Compression Algorithm Based on Mathematical Morphology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |