Video Frame Feature Extraction Using Global and Local Channel Attention

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional video frame feature extraction methods are limited in capturing robust features as they primarily focus on local information, neglecting the impact of temporal context, resulting in insufficient feature extraction for video learning tasks.

Innovation Solution

A method that calculates global channel attention information by fusing initial features from multiple video frames and combines it with local channel attention information to perform channel attention mechanism processing, enhancing feature extraction by integrating both local and global information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional local feature extraction methods are used, then the extraction process is simple and fast, but the robustness of feature extraction is poor

Engineering Contradiction:

Improverobustness of feature extractionVSAvoidcomplexity of feature extraction method

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the feature extraction process into two distinct components: local channel attention (processing individual frame features) and global channel attention (processing temporal relationships across multiple frames). This segmentation allows the system to capture both local details and global context, thereby improving robustness while maintaining manageable complexity through modular processing

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements a nested structure where global channel attention operates on top of local channel attention features. The local features are first extracted and then fed into the global attention mechanism, creating a hierarchical feature representation that enhances robustness by combining multiple levels of abstraction

Inventive Principle:

Principle #7Nested doll (Nesting)

2Measurement precision

If only local features of individual video frames are captured, then the processing speed is fast, but the accuracy of video learning tasks is insufficient

Engineering Contradiction:

Improveaccuracy of video learning tasksVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by selectively attending to important features rather than processing all features uniformly. The channel attention mechanisms (both local and global) compute attention weights to highlight salient features and suppress irrelevant ones, achieving accurate feature representation with reduced processing overhead compared to exhaustive feature analysis

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20240290096A1Video frame feature extraction method, device and computer-readable storage medium

Publication Date: 2024.08.29 UBTECH ROBOTICS CORP LTD

AI summary

A method for extracting video frame features includes: obtaining a number of initial features of each video frame in a video sequence; calculating global channel attention information of the video sequence based on the initial features of each video frame in the video sequence; calculating local channel attention information of a target video frame according to initial features of a target video frame; wherein the target video frame is one of the video frames in the video sequence; and performing channel attention mechanism processing on the initial features of the target video frame according to the global channel attention information and the local channel attention information to obtain optimized features of the target video frame.