Deformable DETR: Deformable Transformers for End-to-End Object Detection

10/08/2020
by   Xizhou Zhu, et al.
10

DETR has been recently proposed to eliminate the need for many hand-designed components in object detection while demonstrating good performance. However, it suffers from slow convergence and limited feature spatial resolution, due to the limitation of Transformer attention modules in processing image feature maps. To mitigate these issues, we proposed Deformable DETR, whose attention modules only attend to a small set of key sampling points around a reference. Deformable DETR can achieve better performance than DETR (especially on small objects) with 10× less training epochs. Extensive experiments on the COCO benchmark demonstrate the effectiveness of our approach. Code shall be released.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
02/18/2022

Task Specific Attention is one more thing you need for object detection

Various models have been proposed to solve the object detection problem....
research
03/28/2022

MSTR: Multi-Scale Transformer for End-to-End Human-Object Interaction Detection

Human-Object Interaction (HOI) detection is the task of identifying a se...
research
01/13/2022

TransVOD: End-to-end Video Object Detection with Spatial-Temporal Transformers

Detection Transformer (DETR) and Deformable DETR have been proposed to e...
research
11/29/2021

Sparse DETR: Efficient End-to-End Object Detection with Learnable Sparsity

DETR is the first end-to-end object detector using a transformer encoder...
research
03/02/2023

FeatAug-DETR: Enriching One-to-Many Matching for DETRs with Feature Augmentation

One-to-one matching is a crucial design in DETR-like object detection fr...
research
01/19/2021

Fast Convergence of DETR with Spatially Modulated Co-Attention

The recently proposed Detection Transformer (DETR) model successfully ap...
research
05/23/2021

End-to-End Video Object Detection with Spatial-Temporal Transformers

Recently, DETR and Deformable DETR have been proposed to eliminate the n...

Please sign up or login with your details

Forgot password? Click here to reset