DenseNet: Implementing Efficient ConvNet Descriptor Pyramids
This addresses runtime and training speed issues for object detection in computer vision, but it is incremental as it builds on existing CNN classifier topologies.
The paper tackles the computational inefficiency of object detection with CNNs by introducing DenseNet, a system that computes dense, multiscale features from convolutional layers to share work across overlapping regions, though no concrete performance numbers are provided.
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.