End-to-End Object Detection with Transformers

TL;DR

DETR introduces a transformer-based approach to object detection that simplifies the pipeline by eliminating hand-designed components and directly predicting object sets with global reasoning, achieving competitive accuracy.

Contribution

The paper presents a novel end-to-end object detection framework using transformers, removing traditional components like NMS and anchors, and demonstrating competitive performance on COCO.

Findings

Achieves accuracy comparable to Faster R-CNN on COCO

Simplifies detection pipeline by removing hand-designed components

Easily extends to panoptic segmentation

Abstract

We present a new method that views object detection as a direct set prediction problem. Our approach streamlines the detection pipeline, effectively removing the need for many hand-designed components like a non-maximum suppression procedure or anchor generation that explicitly encode our prior knowledge about the task. The main ingredients of the new framework, called DEtection TRansformer or DETR, are a set-based global loss that forces unique predictions via bipartite matching, and a transformer encoder-decoder architecture. Given a fixed small set of learned object queries, DETR reasons about the relations of the objects and the global image context to directly output the final set of predictions in parallel. The new model is conceptually simple and does not require a specialized library, unlike many other modern detectors. DETR demonstrates accuracy and run-time performance on par with the well-established and highly-optimized Faster RCNN baseline on the challenging COCO object detection dataset. Moreover, DETR can be easily generalized to produce panoptic segmentation in a unified manner. We show that it significantly outperforms competitive baselines. Training code and pretrained models are available at https://github.com/facebookresearch/detr.