Focal Loss for Dense Object Detection

Tsung-Yi LinPriya GoyalRoss GirshickKaiming HePiotr Dollár

   Papers with code   Abstract  PDF

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)

Benchmarked Models

RANK
MODEL
REPO
CODE RESULT
PAPER RESULT
ε-REPRODUCED
BUILD
1
RetinaNet
(ResNeXt-101 64x4d-FPN…)
0.400
--
2
RetinaNet
(ResNeXt-101 64x4d-FPN…)
0.396
--
3
RetinaNet
(ResNeXt-101 32x4d-FPN…)
0.393
--
4
RetinaNet
(ResNeXt-101 32x4d-FPN…)
0.390
--
5
RetinaNet
(ResNet-101-FPN+, casc…)
0.381
--
6
RetinaNet
(ResNet-101-FPN, 1x LR)
0.377
--
7
RetinaNet
(ResNet-50-FPN, 20e LR)
0.364
--
8
RetinaNet
(ResNet-50-FPN+)
0.356
--