DeepID-Net: Deformable Deep Convolutional Neural Networks for Object Detection

TL;DR

This paper introduces deformable deep convolutional neural networks with a new def-pooling layer and pre-training strategy, significantly improving object detection accuracy over previous methods like RCNN and GoogLeNet.

Contribution

It presents a novel deformable deep learning architecture with def-pooling and a new pre-training approach, enhancing object detection performance and model diversity.

Findings

Improved mean average precision from 31% to 50.3% on ILSVRC2014.

Outperformed GoogLeNet by 6.1% in detection accuracy.

Provided detailed analysis of components for better understanding of the detection pipeline.

Abstract

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 \cite{girshick2014rich}, which was the state-of-the-art, from 31\% to 50.3\% on the ILSVRC2014 detection test set. It also outperforms the winner of ILSVRC2014, GoogLeNet, by 6.1\%. Detailed component-wise analysis is also provided through extensive experimental evaluation, which provide a global view for people to understand the deep learning object detection pipeline.