Video Compressed Sensing Method Based on Feature Awareness
A video compression and video technology, applied in the field of video processing, can solve the problems of reconstruction result distortion, original video deviation, lack, etc., and achieve the effect of enhancing temporal and spatial correlation and good visual effect.
Active Publication Date: 2022-07-26
XIDIAN UNIV
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The perceptual loss is based on the distance between the feature maps extracted from the classification network between the original signal and the reconstructed signal, which is more in line with the visual habits of the human eye, but due to the lack of pixel-level constraints, it is easy to distort the reconstruction result, which is different from the original video. large deviation
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0035] The present embodiment and effects will be further described in detail below with reference to the accompanying drawings.
[0036] refer to figure 1 , the implementation steps are as follows:
[0037] Step 1, prepare the dataset.
[0038] 1a) Install Python on the computer, and install the third-party library cv2 under Python;
[0039] 1b) Download the Myanmar dataset, which contains 59 color videos in different scenes with a resolution of 4K, randomly select 54 videos, and use the video frame conversion function cv2.VideoCapture() to convert them into color video frames , and then crop it into small pieces of 240×240 according to the spatial position, save the small pieces of each spatial position as pictures, store them in different subfolders, and name them according to the chronological order of the video frames, such as 1.jpg, 2 .jpg......, all subfolders constitute the training dataset;
[0040] 1c) Take the randomly downloaded video "basketball" and the remai...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a video compression sensing method based on feature perception, which mainly solves the problems of weak spatiotemporal correlation and poor visual effect of reconstructed video in the prior art. Its implementation plan is: prepare training data set and test data set; design the structure of spatiotemporal video compressed sensing network; design the calculation method of fine sensing loss; write network structure file, network training file and fine sensing loss according to the designed network structure and fine sensing loss. Model test file; use the network training file and training data set to train the network to obtain a trained network model; use the model test file and test data set to process the test video to obtain the corresponding reconstructed video. The spatiotemporal video compressed sensing network trained by the fine perception loss designed by the present invention can enhance the spatiotemporal correlation of the reconstructed video, improve the reconstruction quality, and can generate reconstructed videos with better visual effects, which can be used to achieve better visual effects Compressed sensing reconstruction of video.
Description
technical field [0001] The invention belongs to the technical field of video processing, and mainly relates to a video compressive sensing method, which can be used for realizing video compressive sensing reconstruction with better visual effect. Background technique [0002] Video compressed sensing is a technology that can compress scene video while sampling, which can greatly reduce data storage space and transmission bandwidth. According to different observation methods, the existing video compressive sensing methods can be divided into spatial video compressive sensing and temporal video compressive sensing. The observation process of spatial video compressed sensing is realized by spatial multiplexing cameras, and the input video is observed and reconstructed separately. The temporal video compressive sensing method uses temporal multiplexing cameras to observe the scene, and simultaneously observes and reconstructs multiple consecutive frames of the input video. [...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): H04N19/186H04N19/176H04N19/172H04L41/14G06N3/04
Inventor 谢雪梅刘婉赵至夫石光明
Owner XIDIAN UNIV