TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
Object Counting	CARPK	RetinaNet (2018)	MAE	24.58	# 5
Object Counting	CARPK	RetinaNet (2018)	RMSE	33.12	# 3
Real-Time Object Detection	COCO	FPN R-101 @256	MAP	37.8	# 6
Real-Time Object Detection	COCO	FPN R-101 @256	APM	41.1	# 1
Region Proposal	COCO test-dev	RPN+Focal Loss	AR100	50.2	# 3
Region Proposal	COCO test-dev	RPN+Focal Loss	AR1000	60.9	# 3
Region Proposal	COCO test-dev	RPN+Focal Loss	ARL	67.5	# 3
Region Proposal	COCO test-dev	RPN+Focal Loss	ARM	58.2	# 3
Region Proposal	COCO test-dev	RPN+Focal Loss	ARS	33.9	# 2
Region Proposal	COCO test-dev	RPN+Focal Loss	AR300	56.6	# 2
Object Detection	COCO test-dev	RetinaNet (ResNet-101-FPN)	box AP	39.1	# 59
Object Detection	COCO test-dev	RetinaNet (ResNet-101-FPN)	AP50	59.1	# 56
Object Detection	COCO test-dev	RetinaNet (ResNet-101-FPN)	AP75	42.3	# 56
Object Detection	COCO test-dev	RetinaNet (ResNet-101-FPN)	APS	21.8	# 57
Object Detection	COCO test-dev	RetinaNet (ResNet-101-FPN)	APM	42.7	# 50
Object Detection	COCO test-dev	RetinaNet (ResNet-101-FPN)	APL	50.2	# 58
Object Detection	COCO test-dev	RetinaNet (ResNeXt-101-FPN)	box AP	40.8	# 48
Object Detection	COCO test-dev	RetinaNet (ResNeXt-101-FPN)	AP50	61.1	# 46
Object Detection	COCO test-dev	RetinaNet (ResNeXt-101-FPN)	AP75	44.1	# 48
Object Detection	COCO test-dev	RetinaNet (ResNeXt-101-FPN)	APS	24.1	# 43
Object Detection	COCO test-dev	RetinaNet (ResNeXt-101-FPN)	APM	44.2	# 44
Object Detection	COCO test-dev	RetinaNet (ResNeXt-101-FPN)	APL	51.2	# 53
Dense Object Detection	SKU-110K	RetinaNet	AP	.455	# 3
Dense Object Detection	SKU-110K	RetinaNet	AP75	.389	# 2
Face Identification	Trillion Pairs Dataset	F-Softmax	Accuracy	39.80	# 5
Face Verification	Trillion Pairs Dataset	F-Softmax	Accuracy	37.14	# 5

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/focal-loss-for-dense-object-detection/region-proposal-on-coco-test-dev)](https://paperswithcode.com/sota/region-proposal-on-coco-test-dev?p=focal-loss-for-dense-object-detection)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/focal-loss-for-dense-object-detection/dense-object-detection-on-sku-110k)](https://paperswithcode.com/sota/dense-object-detection-on-sku-110k?p=focal-loss-for-dense-object-detection)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/focal-loss-for-dense-object-detection/object-counting-on-carpk)](https://paperswithcode.com/sota/object-counting-on-carpk?p=focal-loss-for-dense-object-detection)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/focal-loss-for-dense-object-detection/face-identification-on-trillion-pairs-dataset)](https://paperswithcode.com/sota/face-identification-on-trillion-pairs-dataset?p=focal-loss-for-dense-object-detection)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/focal-loss-for-dense-object-detection/face-verification-on-trillion-pairs-dataset)](https://paperswithcode.com/sota/face-verification-on-trillion-pairs-dataset?p=focal-loss-for-dense-object-detection)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/focal-loss-for-dense-object-detection/real-time-object-detection-on-coco)](https://paperswithcode.com/sota/real-time-object-detection-on-coco?p=focal-loss-for-dense-object-detection)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/focal-loss-for-dense-object-detection/object-detection-on-coco)](https://paperswithcode.com/sota/object-detection-on-coco?p=focal-loss-for-dense-object-detection)`

Focal Loss for Dense Object Detection

ICCV 2017 • Tsung-Yi Lin • Priya Goyal • Ross Girshick • Kaiming He • Piotr Dollár

The highest accuracy object detectors to date are based on a two-stage approach popularized by R-CNN, where a classifier is applied to a sparse set of candidate object locations. In contrast, one-stage detectors that are applied over a regular, dense sampling of possible object locations have the potential to be faster and simpler, but have trailed the accuracy of two-stage detectors thus far... (read more)

PDF Abstract

Code

Add Remove

tensorflow/models

64,009

facebookresearch/detectron

23,271

AlexeyAB/darknet

↳ Quickstart in

Colab

10,079

fizyr/keras-retinanet

3,759

tryolabs/luminoth

2,301

See all 143 implementations

Tasks

Add Remove

DENSE OBJECT DETECTION

OBJECT DETECTION

REAL-TIME OBJECT DETECTION

REGION PROPOSAL

Results from the Paper

Edit

#3 best model for Dense Object Detection on SKU-110K

Get a GitHub badge

TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK	LEADERBOARD
Real-Time Object Detection	COCO	FPN R-101 @256	MAP	37.8	# 6	Compare
Real-Time Object Detection	COCO	FPN R-101 @256	APM	41.1	# 1	Compare
Region Proposal	COCO test-dev	RPN+Focal Loss	AR100	50.2	# 3	Compare
			AR1000	60.9	# 3	Compare
			ARL	67.5	# 3	Compare
			ARM	58.2	# 3	Compare
			ARS	33.9	# 2	Compare
			AR300	56.6	# 2	Compare
Object Detection	COCO test-dev	RetinaNet (ResNet-101-FPN)	box AP	39.1	# 59	Compare
			AP50	59.1	# 56	Compare
			AP75	42.3	# 56	Compare
			APS	21.8	# 57	Compare
			APM	42.7	# 50	Compare
			APL	50.2	# 58	Compare
Object Detection	COCO test-dev	RetinaNet (ResNeXt-101-FPN)	box AP	40.8	# 48	Compare
			AP50	61.1	# 46	Compare
			AP75	44.1	# 48	Compare
			APS	24.1	# 43	Compare
			APM	44.2	# 44	Compare
			APL	51.2	# 53	Compare

Results from Other Papers

TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK	COMPARE
Object Counting	CARPK	RetinaNet (2018)	MAE	24.58	# 5	See all
Object Counting	CARPK	RetinaNet (2018)	RMSE	33.12	# 3	See all
Dense Object Detection	SKU-110K	RetinaNet	AP	.455	# 3	See all
Dense Object Detection	SKU-110K	RetinaNet	AP75	.389	# 2	See all
Face Identification	Trillion Pairs Dataset	F-Softmax	Accuracy	39.80	# 5	See all
Face Verification	Trillion Pairs Dataset	F-Softmax	Accuracy	37.14	# 5	See all

Focal Loss for Dense Object Detection

Code Edit Add Remove

Tasks Edit Add Remove

Results from the Paper Edit

Results from Other Papers

Code

Add Remove

Tasks

Add Remove

Results from the Paper

Edit