Video Feature Extraction Using Motion Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video feature extraction methods consume large storage space and decoding time, and the slicing methods introduce redundant computations, reducing efficiency.
Innovation Solution
A method that determines groups of pictures in video data, extracts frame features from intra coding frames and predictive-frames, and updates compensation features to obtain accurate frame features, thereby reducing storage and decoding requirements and avoiding redundant computations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video data is decoded into multiple video frames for feature extraction, then feature extraction accuracy is improved, but storage space consumption increases
Solution Approach 1:
The patent extracts only the essential feature information from video frames rather than storing and processing complete decoded frames. By extracting features such as motion vectors, key points, and semantic information, the system reduces storage requirements while maintaining feature extraction accuracy for video slicing tasks.
Solution Approach 2:
The patent segments the video data into groups of pictures (GOPs) and processes features at the frame level rather than handling entire decoded video streams. This segmentation allows selective extraction and storage of only necessary feature data, reducing overall storage space while preserving accuracy for boundary detection.
2Measurement precision
If video frames are decoded and features are extracted for slicing, then slicing accuracy is improved, but decoding time increases
Solution Approach 1:
The patent performs preliminary feature extraction on intra-coding frames and motion compensation information during the encoding process rather than waiting for complete decoding. This preliminary action enables the system to prepare feature data in advance, reducing the time required for subsequent slicing operations while maintaining accuracy.
Solution Approach 2:
The patent introduces motion compensation information as an intermediary element that bridges the gap between encoded data and final frame features. By using motion compensation information to reconstruct frame features without fully decoding the video, the system achieves slicing accuracy while significantly reducing decoding time.
3Measurement precision
If features are extracted from all input frames for boundary prediction, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies local quality by extracting and processing features selectively based on the specific requirements of boundary detection. Instead of uniformly extracting features from all frames, the system focuses computational resources on frames and regions where boundary information is most relevant, reducing overall computational complexity while maintaining prediction accuracy.
Solution Approach 2:
The patent uses partial action by extracting features from a subset of critical frames (such as intra-coding frames and key predictive frames) rather than all frames. This partial feature extraction approach provides sufficient information for accurate boundary prediction while significantly reducing the computational burden compared to processing every frame.
4Quantity of substance
If redundant feature extraction is performed during slicing, then feature completeness is improved, but processing efficiency decreases
Solution Approach 1:
The patent discards redundant feature extraction operations by leveraging motion compensation information to derive frame features directly from encoded data. Instead of re-extracting features from all input frames, the system recovers necessary feature information through motion compensation, eliminating redundant computations while maintaining feature completeness for accurate slicing.
Data Source
AI summary
A feature extraction method and apparatus for a video, a slicing method and apparatus for a video, and an electronic device and a storage medium. The method includes: determining a plurality of groups of pictures of video data, in which each group of pictures includes, according to a time sequence, an intra coding frame and at least one predictive-frame; for each group of pictures, extracting a first frame feature of the intra coding frame, and extracting a compensation feature of motion compensation information of the at least one predictive-frame relative to the intra coding frame; and updating the compensation feature according to the first frame feature to obtain a second frame feature of the at least one predictive-frame, so as to obtain a frame feature of a video frame in the video data.