Data Science Talk Series 

Title: Visual Object Tracking: Benchmarks & Algorithms

Abstract:

Visual tracking has been one of the most fundamental problems in computer vision. In recent years, great progress has been witnessed in the tracking community owing to the proposal of many tracking benchmarks and algorithms, particularly in the deep learning era. In this talk, I will discuss my research on visual tracking from two aspects: benchmarks and algorithms. On one side, I will talk about my works on various benchmarks for single object tracking, multi-object tracking, and some special topics. On the other side, I will discuss one of our recent works on scaling model size for improving tracking. Focus will be on how to efficiently train larger model models for improving tracking performance. I hope this research will inspire more exploration on both tracking benchmarks and algorithms in the future research. All data and code in this talk can be found at my website: https://hengfan2010.github.io/.