Object Feature Extraction via Quality Index Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object feature extraction methods fail to consider motion blur, occlusion, and varying illumination conditions, leading to suboptimal feature extraction and matching accuracy, especially when processing videos in real-time with limited resources.
Innovation Solution
An object feature extraction device that calculates a quality index based on size, motion, posture, and environmental factors to selectively prioritize feature extraction for objects, using a combination of object detection, tracking, and storage mechanisms to predict feature quality and adaptively select objects for extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If features are extracted from all objects appearing on the image, then matching accuracy is improved, but processing time increases and real-time processing becomes difficult
Solution Approach 1:
The patent applies partial action by selecting and extracting features only from specific objects that meet certain criteria (quality index threshold) rather than processing all objects. This reduces the number of feature extraction operations while maintaining matching accuracy for relevant objects, thereby decreasing overall processing time and enabling real-time video analysis.
Solution Approach 2:
The patent implements local quality by evaluating and selecting objects based on their individual quality indices, which are calculated from multiple factors including size, motion, posture, and environmental conditions. Each object is assessed locally according to its specific characteristics, allowing the system to prioritize feature extraction for objects most likely to yield useful matching results while skipping others.
2Productivity
If feature extraction is performed on each frame with limited calculation resources, then processing speed is maintained, but feature extraction quality deteriorates
Solution Approach 1:
The system performs feature extraction only for objects that meet the quality criteria in selected frames rather than processing all objects in every frame. This partial processing approach maintains processing speed by reducing the total number of extraction operations while improving feature quality by focusing resources on objects with favorable extraction conditions.
Solution Approach 2:
The patent calculates quality indices for objects before performing feature extraction, using preliminary assessment of size, motion, posture, and environmental factors. This preliminary action allows the system to pre-identify which objects are worth processing, ensuring that feature extraction resources are allocated to objects most likely to produce high-quality features, thereby maintaining both speed and quality.
3Device complexity
If objects are selected based only on size for feature extraction, then processing load is reduced, but feature quality deteriorates due to motion blur and other factors
Solution Approach 1:
The patent changes the selection parameters from a single criterion (size only) to multiple parameters including size, motion amount, posture, and environmental factors like illumination and occlusion. This multi-parameter quality index calculation provides a more comprehensive assessment of feature extraction suitability, improving feature quality while maintaining reasonable processing load by still selecting only a subset of objects.
Solution Approach 2:
The system performs preliminary evaluation of multiple quality factors (size, motion, posture, environment) before feature extraction to identify objects with favorable conditions. This preliminary multi-factor assessment ensures that selected objects are likely to produce high-quality features, compensating for the reduced number of processed objects and maintaining overall feature quality despite lower processing load.
4Measurement precision
If feature extraction frequency is increased for objects in the lower region of the image, then matching accuracy for moving objects is improved, but processing time increases
Solution Approach 1:
The patent applies local quality assessment by calculating quality indices individually for each object based on its specific position, motion characteristics, size, and environmental conditions. Objects in the lower region with high motion are not automatically prioritized; instead, each object is evaluated locally, and feature extraction is performed only for those meeting the quality threshold, thus improving matching accuracy for relevant objects without uniformly increasing processing time for all lower-region objects.
Data Source
AI summary
According to an example embodiment, an object feature extraction device includes a video acquisition means for acquiring a video and generating the acquired video as an image sequence, an object detection means for detecting an object from the generated image and generating a detection result, an object tracking means for tracking the object based on the generated image and the detection result and generating a tracking result, an image storage means for storing the image, a detection result storage means for storing the detection result, a tracking result storage means for storing the tracking result, an object selection means for calculating, based on the detection result and the tracking result, a quality index for predicting the quality of a feature of the object detected at an extraction time, selecting the object to be subjected to feature extraction, and generating object selection information.