Object Detection Made Simpler by Eliminating Heuristic NMS

TL;DR

This paper introduces an NMS-free, end-to-end object detection framework that simplifies the detection pipeline by removing the need for non-maximum suppression, achieving comparable or better accuracy with minimal modifications.

Contribution

The authors propose a minimal modification to existing one-stage detectors to eliminate NMS, using a compact PSS head and stop-gradient training to enable single positive sample selection per object.

Findings

Achieves comparable or improved accuracy on COCO dataset.

Eliminates the need for NMS during inference, maintaining similar speed.

Outperforms previous end-to-end NMS-free detectors.

Abstract

We show a simple NMS-free, end-to-end object detection framework, of which the network is a minimal modification to a one-stage object detector such as the FCOS detection model [Tian et al. 2019]. We attain on par or even improved detection accuracy compared with the original one-stage detector. It performs detection at almost the same inference speed, while being even simpler in that now the post-processing NMS (non-maximum suppression) is eliminated during inference. If the network is capable of identifying only one positive sample for prediction for each ground-truth object instance in an image, then NMS would become unnecessary. This is made possible by attaching a compact PSS head for automatic selection of the single positive sample for each instance (see Fig. 1). As the learning objective involves both one-to-many and one-to-one label assignments, there is a conflict in the labels of some training examples, making the learning challenging. We show that by employing a stop-gradient operation, we can successfully tackle this issue and train the detector. On the COCO dataset, our simple design achieves superior performance compared to both the FCOS baseline detector with NMS post-processing and the recent end-to-end NMS-free detectors. Our extensive ablation studies justify the rationale of the design choices.