Kernelized Correlation Filter Multi-Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Recognition-based tracking algorithms are time-consuming and struggle with multi-scale problems, especially when dealing with a large number of training samples, and are not suitable for multi-object tracking scenarios.
Innovation Solution
The high-speed automatic multi-object tracking method employs kernelized correlation filters, using a circulant matrix to obtain training samples and accelerate the training process with faster Fourier transform, and performs multi-scale analysis via optical flow to efficiently track objects across multiple frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If recognition-based tracking algorithms use online machine learning to train classifiers, then tracking robustness and adaptability improve, but training time increases significantly
Solution Approach 1:
The patent pre-generates a large number of training samples using circulant matrices before the actual tracking process. This preliminary sample generation allows the classifier to be trained in advance with sufficient data, improving tracking robustness while enabling faster real-time tracking since the heavy sample generation work is completed beforehand.
Solution Approach 2:
The patent replaces traditional mechanical training processes with FFT-based frequency domain computation. By transforming the training process into the frequency domain using Fast Fourier Transform, the computational complexity is reduced from O(N²) to O(N log N), dramatically accelerating the training process while maintaining classification accuracy.
2Measurement precision
If circulant matrix method is used to generate training samples, then sufficient training samples are obtained with higher recognition rate, but multi-scale problems cannot be solved
Solution Approach 1:
The patent introduces dynamic scaling factors and multi-scale feature extraction mechanisms that allow the tracking system to adapt to objects of different sizes. The circulant matrix method is extended to work with scaled versions of training samples, enabling the classifier to recognize objects at multiple scales while maintaining high recognition rates.
3Loss of time
If traditional Fourier transform is used for kernel trick acceleration, then training time is reduced, but computational efficiency is not maximized
Solution Approach 1:
The patent employs Fast Fourier Transform (FFT) to replace traditional Fourier transform computation. This substitution leverages the efficient O(N log N) algorithm of FFT, significantly reducing computational time and improving processing speed while maintaining the accuracy benefits of the kernel trick for classification.
4Measurement precision
If recognition-based tracking is applied to multi-object scenarios, then individual object tracking accuracy is maintained, but the system cannot be extended to handle multiple objects efficiently
Solution Approach 1:
The patent segments the multi-object tracking problem into independent sub-problems by generating separate circulant matrix training samples for each object class. This segmentation allows the system to maintain high tracking accuracy for each individual object while efficiently handling multiple objects through parallel processing of their respective training sets and classifiers.
Data Source
AI summary
A high-speed automatic multi-object tracking method with kernelized correlation filters is provided. The method includes obtaining an image frame from a plurality of image frames in a video, extracting a foreground object sequence from the obtained image frame, and determining similarity between each foreground object of the extracted foreground object sequence and a tracked object. The method also includes calculating HOG features of the foreground objects with a lower similarity, obtaining training samples for each of the foreground objects with the lower similarity using a circular matrix, obtaining a classifier via a kernel method accelerated by FFTW, and obtaining tracking points using a sparse optical flow. Further, the method includes detecting object matching responses using a detection response function, performing multi-scale analysis for the object based on an optical flow method, and processing a next image frame of the plurality of image frames in the video until the video ends.