Deep Residual Learning for Image Recognition

CVPR 2016 Kaiming He • Xiangyu Zhang • Shaoqing Ren • Jian Sun

We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We also present analysis on CIFAR-10 with 100 and 1000 layers. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Full paper



Task Dataset Model Metric name Metric value Global rank Compare
Image Classification CIFAR-10 ResNet Percentage correct 93.57 # 19
Image Classification CIFAR-10 ResNet Percentage error 6.43 # 19
Object Detection COCO Faster R-CNN + box refinement + context + multi-scale testing Bounding Box AP 34.9 # 18
Object Detection PASCAL VOC 2007 ResNet-101 MAP 76.4% # 10