DetNet: Design Backbone for Object Detection

ECCV 2018 • Zeming Li • Chao Peng • Gang Yu • Xiangyu Zhang • Yangdong Deng • Jian Sun

Recent CNN based object detectors, either one-stage methods like YOLO, SSD, and RetinaNet, or two-stage detectors like Faster R-CNN, R-FCN and FPN, are usually trying to directly finetune from ImageNet pre-trained models designed for the task of image classification. However, there has been little work discussing the backbone feature extractor specifically designed for the task of object detection... (read more)

PDF Abstract

Results from the Paper

Add Remove

Submit results from this paper to get state-of-the-art GitHub badges and help the community compare results to other papers.

METHOD	TYPE
RPN	Region Proposal
Dilated Convolution	Convolutions
ReLU	Activation Functions
RoIPool	RoI Feature Extractors
Softmax	Output Functions
Residual Connection	Skip Connections
Faster R-CNN	Object Detection Models
Average Pooling	Pooling Operations
Dilated Bottleneck with Projection Block	Skip Connection Blocks
Dilated Bottleneck Block	Skip Connection Blocks
Non Maximum Suppression	Proposal Filtering
Global Average Pooling	Pooling Operations
SSD	Object Detection Models
DetNet	Convolutional Neural Networks
Focal Loss	Loss Functions
1x1 Convolution	Convolutions
FPN	Feature Extractors
RetinaNet	Object Detection Models
Position-Sensitive RoI Pooling	RoI Feature Extractors
Convolution	Convolutions
R-FCN	Object Detection Models

METHOD

TYPE

RPN