Research Article
A Deconvolutional Strategy for Implementing Large Patch Sizes Supports Improved Image Classification
@INPROCEEDINGS{10.4108/eai.3-12-2015.2262498, author={Xinhua Zhang and Garrett Kenyon}, title={A Deconvolutional Strategy for Implementing Large Patch Sizes Supports Improved Image Classification}, proceedings={The First International Workshop on Computational Models of the Visual Cortex: Hierarchies, Layers, Sparsity, Saliency and Attention}, publisher={ACM}, proceedings_a={CMVC}, year={2016}, month={5}, keywords={sparse coding unsupervised feature learning whole image classification}, doi={10.4108/eai.3-12-2015.2262498} }- Xinhua Zhang
Garrett Kenyon
Year: 2016
A Deconvolutional Strategy for Implementing Large Patch Sizes Supports Improved Image Classification
CMVC
ACM
DOI: 10.4108/eai.3-12-2015.2262498
Abstract
Sparse coding is a widely-used technique for learning an overcomplete basis set from unlabeled image data. We hypothesize that as the size of the image patch spanned by each basis vector increases, the resulting dictionary should encompass a broader range of spatial scales, including more features that better discriminate between object classes. Previous efforts to measure the effects of patch size on image classification performance were confounded by the difficulty of maintaining a given level of overcompleteness as the patch size is increased. Here, we employ a type of deconvolutional network in which overcompleteness is independent of patch size. Based on image classification results on the CIFAR10 database, we find that optimizing our deconvolutional network for sparse reconstruction leads to improved classification performance as a function of the number of training epochs. Different from previous reports, we find that enforcing a certain degree of sparsity improves classification performance. We also find that classification performance improves as both the number of learned features (dictionary size) and the size of the image patch spanned by each feature (patch size) are increased, ultimately the best published results for sparse autoencoders