DeepID-Net: Object Detection with Deformable Part Based Convolutional Neural Networks

In this paper, we propose deformable deep convolutional neural networks for generic object detection. This new deep learning object detection framework has innovations in multiple aspects. In the proposed new deep architecture, a new deformation constrained pooling (def-pooling) layer models the def...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on pattern analysis and machine intelligence 2017-07, Vol.39 (7), p.1320-1334
Main Authors: Ouyang, Wanli, Zeng, Xingyu, Wang, Xiaogang, Qiu, Shi, Luo, Ping, Tian, Yonglong, Li, Hongsheng, Yang, Shuo, Wang, Zhe, Li, Hongyang, Wang, Kun, Yan, Junjie, Loy, Chen-Change, Tang, Xiaoou
Format: Article
Language:English
Subjects:
Cameras
CNN
Computer architecture
Context modeling
convolutional neural networks
deep learning
deep model
Deformable models
Deformation
Formability
Geometric constraints
Infrared imaging systems
Learning
Machine learning
Neural networks
Object detection
Object recognition
Representations
State of the art
Training
Visualization
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we propose deformable deep convolutional neural networks for generic object detection. This new deep learning object detection framework has innovations in multiple aspects. In the proposed new deep architecture, a new deformation constrained pooling (def-pooling) layer models the deformation of object parts with geometric constraint and penalty. A new pre-training strategy is proposed to learn feature representations more suitable for the object detection task and with good generalization capability. By changing the net structures, training strategies, adding and removing some key components in the detection pipeline, a set of models with large diversity are obtained, which significantly improves the effectiveness of model averaging. The proposed approach improves the mean averaged precision obtained by RCNN [1] , which was the state-of-the-art, from 31 to 50.3 percent on the ILSVRC2014 detection test set. It also outperforms the winner of ILSVRC2014, GoogLeNet, by 6.1 percent. Detailed component-wise analysis is also provided through extensive experimental evaluation, which provides a global view for people to understand the deep learning object detection pipeline.
ISSN:0162-8828
1939-3539
2160-9292
DOI:10.1109/TPAMI.2016.2587642