Efficient Hybrid DCT-Wiener Algorithm Based Deep Learning Approach For Semantic Shape Segmentation

Abstract

    Semantic segmentation is effective in numerous object classification tasks such as autonomous vehicles and scene understanding. With the advent in the deep learning domain, lots of efforts are seen in applying deep learning algorithms for semantic segmentation. Most of the algorithms gain the required accuracy while compromising on their storage and computational requirements. The work showcases the implementation of Convolutional Neural Network (CNN) using Discrete Cosine Transform (DCT), where DCT exhibit exceptional energy compaction properties. The proposed Adaptive Weight Wiener Filter (AWWF) rearranges the DCT coefficients by truncating the high frequency coefficients. AWWF-DCT model reinstate the convolutional layers giving modularity in the design using multi scale convolution block. The impact of selection of DCT coefficients in the proposed model is validated on the benchmark database as City Spaces. The same level of accuracy compared to the conventional algorithm is achieved using only 40 % of the DCT coefficients. Extensive experiments validate the advantages of adaptive DCT modeling of CNN in semantic segmentation and image classification.