Software for Visual Insect Tracking Based on F-transform Pattern Matching

摘要

We introduce a problem of tracking small animals, especially insects. To solve this problem, we focus on visual tracking in recorded movies, propose our pattern tracking mechanism based on F -transform, and implement a user-friendly software to handle the movies. The tracking core is compared with five state-of-the-art tracking algorithms: KCF, MIL, TLD, Boosting and MedianFlow from processing time and algorithm failure rate point of views. Based on the results computed from 1000 movie frames, we observed that the proposed F-transform tracking core is the fastest and the most reliable method.