The 1st Tiny Object Detection Challenge:Methods and Results

Abstract

The 1st Tiny Object Detection (TOD) Challenge aims to encourage research in developing novel and accurate methods for tiny object detection in images which have wide views, with a current focus on tiny person detection. The TinyPerson dataset was used for the TOD Challenge and is publicly released. It has 1610 images and 72651 box-level annotations. Around 36 participating teams from the globe competed in the 1st TOD Challenge. In this paper, we provide a brief summary of the 1st TOD Challenge including brief introductions to the top three methods.The submission leaderboard will be reopened for researchers that are interested in the TOD challenge. The benchmark dataset and other information can be found at: https://github.com/ucas-vg/TinyBenchmark.

PAFNet: An Efficient Anchor-Free Object Detector Guidance

“…Training Details For server side, we use ResNet50vd [46] as the backbone network for our ablation study, which is initialized with the weights distilled on Ima-geNet [7] dataset and performs better than ResNet50 [1]. To speed up training and inference, we resize images to 512 × 512, which is slightly different from the common setting.…”