Parallel Particle Filter Tracking for Video Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Real-time object tracking in video data is hindered by high computation workload and issues like partial occlusion and illumination variances, making it difficult to accurately track objects.
Innovation Solution
The implementation of a parallel particle filter tracking method using multiple information cues, such as color histograms, edge orientation, and wavelet features, performed across multiple processing units including CPUs and GPUs, to improve computational speed and address tracking challenges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional object tracking methods are used, then tracking accuracy can be maintained, but computation time increases and real-time tracking becomes difficult
Solution Approach 1:
The patent divides the particle filter tracking computation into multiple independent parallel processes, where each processing unit handles a subset of particles. This segmentation allows simultaneous computation of weight measurements for different particles without interfering with each other, thereby reducing total computation time while maintaining tracking accuracy through comprehensive particle evaluation.
2Reliability
If multiple information cues are used to improve tracking under occlusion and illumination changes, then tracking reliability improves, but computation workload increases
Solution Approach 1:
The patent segments the computation of multiple information cues (color histograms, edge orientation, wavelet features) across different processing units. Each processing unit computes a portion of the feature extractions and weight measurements in parallel, reducing the overall computation workload while maintaining the benefits of multiple cues for improved tracking reliability under challenging conditions.
Data Source
AI summary
A method for tracking an object in a video data, comprises the steps of determining a plurality of particles for estimating a location of the object in the video data, determining a weight for each of the plurality of the particles, wherein the weights of two or more particles are determined substantially in parallel, and estimating the location of the object in the video data based upon the determined particle weights.