TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	AP	14.8	# 2
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	AP50	32.8	# 3
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	AP75	11.4	# 2
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	APvt	0.0	# 3
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	APt	10.8	# 3
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	APs	28.3	# 2
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	APm	28.0	# 2
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	box mAP	54.7	# 49
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	AP50	73.5	# 12
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	AP75	60.1	# 15
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	APS	37.4	# 9
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	APM	57.3	# 13
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	APL	66.4	# 17
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, single-scale)	box mAP	53.3	# 59
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, single-scale)	AP50	71.6	# 23
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, single-scale)	AP75	58.5	# 22
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, single-scale)	APS	33.9	# 25
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, single-scale)	APM	56.5	# 16
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, single-scale)	APL	66.9	# 13
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	box mAP	55.7	# 44
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	AP50	74.2	# 9
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	AP75	61.1	# 11
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	APS	37.7	# 7
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	APM	58.4	# 11
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	APL	68.1	# 11
Panoptic Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	PQ	50	# 15
Panoptic Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	PQst	37.2	# 19
Panoptic Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	PQth	58.5	# 10
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	mask AP	48.5	# 26
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	AP50	72.0	# 8
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	AP75	53.3	# 7
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	APS	31.6	# 6
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	APM	50.9	# 7
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	APL	61.5	# 9
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	mask AP	47.1	# 31
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	AP50	71.1	# 9
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	AP75	51.6	# 8
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	APS	30.3	# 8
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	APM	49.5	# 9
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	APL	59.6	# 11

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detectors-detecting-objects-with-recursive-1/object-detection-on-ai-tod)](https://paperswithcode.com/sota/object-detection-on-ai-tod?p=detectors-detecting-objects-with-recursive-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detectors-detecting-objects-with-recursive-1/panoptic-segmentation-on-coco-test-dev)](https://paperswithcode.com/sota/panoptic-segmentation-on-coco-test-dev?p=detectors-detecting-objects-with-recursive-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detectors-detecting-objects-with-recursive-1/instance-segmentation-on-coco)](https://paperswithcode.com/sota/instance-segmentation-on-coco?p=detectors-detecting-objects-with-recursive-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detectors-detecting-objects-with-recursive-1/object-detection-on-coco)](https://paperswithcode.com/sota/object-detection-on-coco?p=detectors-detecting-objects-with-recursive-1)`

DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution

CVPR 2021 · Siyuan Qiao, Liang-Chieh Chen, Alan Yuille ·

Many modern object detectors demonstrate outstanding performances by using the mechanism of looking and thinking twice. In this paper, we explore this mechanism in the backbone design for object detection. At the macro level, we propose Recursive Feature Pyramid, which incorporates extra feedback connections from Feature Pyramid Networks into the bottom-up backbone layers. At the micro level, we propose Switchable Atrous Convolution, which convolves the features with different atrous rates and gathers the results using switch functions. Combining them results in DetectoRS, which significantly improves the performances of object detection. On COCO test-dev, DetectoRS achieves state-of-the-art 55.7% box AP for object detection, 48.5% mask AP for instance segmentation, and 50.0% PQ for panoptic segmentation. The code is made publicly available.

PDF Abstract CVPR 2021 PDF CVPR 2021 Abstract

Code

Add Remove Mark official

joe-siyuan-qiao/DetectoRS official

1,128

open-mmlab/mmdetection

27,708

FenHua/Robust_Logo_Detection

Vishal-V/tf-models

↳ Quickstart in

Colab

TeamA2020/Practice

See all 6 implementations

Tasks

Add Remove

Instance Segmentation

Object

object-detection

Object Detection

Panoptic Segmentation

Segmentation

Semantic Segmentation

Datasets

MS COCO AI-TOD

Results from the Paper

Edit

Ranked #2 on Object Detection on AI-TOD

Get a GitHub badge

Task	Dataset	Model	Metric Name	Metric Value	Global Rank	Benchmark
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	box mAP	54.7	# 49	Compare
			AP50	73.5	# 12	Compare
			AP75	60.1	# 15	Compare
			APS	37.4	# 9	Compare
			APM	57.3	# 13	Compare
			APL	66.4	# 17	Compare
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, single-scale)	box mAP	53.3	# 59	Compare
			AP50	71.6	# 23	Compare
			AP75	58.5	# 22	Compare
			APS	33.9	# 25	Compare
			APM	56.5	# 16	Compare
			APL	66.9	# 13	Compare
Object Detection	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	box mAP	55.7	# 44	Compare
			AP50	74.2	# 9	Compare
			AP75	61.1	# 11	Compare
			APS	37.7	# 7	Compare
			APM	58.4	# 11	Compare
			APL	68.1	# 11	Compare
Panoptic Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	PQ	50	# 15	Compare
			PQst	37.2	# 19	Compare
			PQth	58.5	# 10	Compare
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-64x4d, multi-scale)	mask AP	48.5	# 26	Compare
			AP50	72.0	# 8	Compare
			AP75	53.3	# 7	Compare
			APS	31.6	# 6	Compare
			APM	50.9	# 7	Compare
			APL	61.5	# 9	Compare
Instance Segmentation	COCO test-dev	DetectoRS (ResNeXt-101-32x4d, multi-scale)	mask AP	47.1	# 31	Compare
			AP50	71.1	# 9	Compare
			AP75	51.6	# 8	Compare
			APS	30.3	# 8	Compare
			APM	49.5	# 9	Compare
			APL	59.6	# 11	Compare

Results from Other Papers

Task	Dataset	Model	Metric Name	Metric Value	Rank	Compare
Object Detection	AI-TOD	DetectoRS (ResNet-50-FPN)	AP	14.8	# 2	See all
			AP50	32.8	# 3	See all
			AP75	11.4	# 2	See all
			APvt	0.0	# 3	See all
			APt	10.8	# 3	See all
			APs	28.3	# 2	See all
			APm	28.0	# 2	See all

Methods

Add Remove

1x1 Convolution • Average Pooling • Convolution • Dilated Convolution • Global Average Pooling • RFP • SAC • Sigmoid Activation

Edit Social Preview

DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution

Code Edit Add Remove Mark official

Tasks Edit Add Remove

Datasets Edit

Results from the Paper Edit