DCT-Conv: Coding filters in convolutional networks with Discrete Cosine Transform
arXiv:2001.08517
Abstract
Convolutional neural networks are based on a huge number of trained weights. Consequently, they are often data-greedy, sensitive to overtraining, and learn slowly. We follow the line of research in which filters of convolutional neural layers are determined on the basis of a smaller number of trained parameters. In this paper, the trained parameters define a frequency spectrum which is transformed into convolutional filters with Inverse Discrete Cosine Transform (IDCT, the same is applied in decompression from JPEG). We analyze how switching off selected components of the spectra, thereby reducing the number of trained weights of the network, affects its performance. Our experiments show that coding the filters with trained DCT parameters leads to improvement over traditional convolution. Also, the performance of the networks modified this way decreases very slowly with the increasing extent of switching off these parameters. In some experiments, a good performance is observed when even 99.9% of these parameters are switched off.
6 pages, 2 figures, submitted for IJCNN 2020
References in corpus (6)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- Compressing Neural Networks with the Hashing Trick
- Rotation equivariant vector field networks
- Scale-Invariant Convolutional Neural Networks
- LeukoNet: DCT-based CNN architecture for the classification of normal versus Leukemic blasts in B-ALL Cancer