2nd Place Solution to ECCV 2020 VIPriors Object Detection Challenge

TL;DR

This paper presents a competitive object detection system for the ECCV 2020 VIPriors Challenge, achieving second place by leveraging advanced data augmentation, model design, and ensemble methods without pre-training.

Contribution

The authors demonstrate that high-performing object detection is possible with limited data by combining novel augmentation, model architecture, and ensemble techniques.

Findings

Achieved 36.6% AP on COCO 2017 validation set

Performed well without pre-training or transfer learning

Secured 2nd place in the challenge

Abstract

In this report, we descibe our approach to the ECCV 2020 VIPriors Object Detection Challenge which took place from March to July in 2020. We show that by using state-of-the-art data augmentation strategies, model designs, and post-processing ensemble methods, it is possible to overcome the difficulty of data shortage and obtain competitive results. Notably, our overall detection system achieves 36.6$\%$ AP on the COCO 2017 validation set using only 10K training images without any pre-training or transfer learning weights ranking us 2nd place in the challenge.