End-End-Object-Detection-By-Using-Transformers-

This paper presents a method for object detection that addresses the problem of direct set prediction. It eliminates the need for multiple components and simplifies the process by avoiding the need for manual intervention.The main components of a framework known as DEtection TRansformer are the transformer and the matching mechanism. This paper shows that the relationship between the two components is strong enough to explain the relationship between the two datasets,The DETR framework achieves a fast R-CNN baseline and is capable of providing uniform panoptic segmentation. It is also well-designed to improve its performance.