DenseNet: Implementing Efficient ConvNet Descriptor Pyramids

TL;DR

DenseNet introduces an efficient system for computing dense, multiscale CNN features to improve object detection performance, sharing computation among overlapping regions to reduce runtime.

Contribution

It presents DenseNet, an open source system that efficiently computes dense CNN features, enabling faster object detection by sharing computations across overlapping regions.

Findings

Enables shared computation for overlapping regions

Provides dense, multiscale CNN features

Open source implementation available

Abstract

Convolutional Neural Networks (CNNs) can provide accurate object classification. They can be extended to perform object detection by iterating over dense or selected proposed object regions. However, the runtime of such detectors scales as the total number and/or area of regions to examine per image, and training such detectors may be prohibitively slow. However, for some CNN classifier topologies, it is possible to share significant work among overlapping regions to be classified. This paper presents DenseNet, an open source system that computes dense, multiscale features from the convolutional layers of a CNN based object classifier. Future work will involve training efficient object detectors with DenseNet feature descriptors.