Tracking Small Birds by Detection Candidate Region Filtering and Detection History-aware Association

TL;DR

This paper introduces Adaptive SAHI and DHSC methods to improve small bird tracking in panoramic videos by enhancing detection accuracy and association speed, verified on the NUBird2022 dataset.

Contribution

The paper presents novel Adaptive SAHI and DHSC techniques specifically designed for small object tracking, improving accuracy and efficiency over existing methods.

Findings

Enhanced tracking accuracy on NUBird2022 dataset

Improved detection speed for small objects

Better association accuracy across frames

Abstract

This paper focuses on tracking birds that appear small in a panoramic video. When the size of the tracked object is small in the image (small object tracking) and move quickly, object detection and association suffers. To address these problems, we propose Adaptive Slicing Aided Hyper Inference (Adaptive SAHI), which reduces the candidate regions to apply detection, and Detection History-aware Similarity Criterion (DHSC), which accurately associates objects in consecutive frames based on the detection history. Experiments on the NUBird2022 dataset verifies the effectiveness of the proposed method by showing improvements in both accuracy and speed.