Quality Improvement of Dot Diffused Block Truncation Coding using Convolutional Neural Networks
Heri Prasetyo, Alim Wicaksono Hari Prayuda, Jing-Ming Guo · 2021
This paper presents a simple technique for improving the quality of Dot Diffused Block Truncation (DDBTC) decoded image. The proposed method is designed based on the Convolutional Neural Network (CNN) framework. It consists of several layers to suppress the impulsive noise, the blocking artifact, and false contours occurred in the DDBTC decoded image. The proposed framework works in the end-to-end learning for the quality enhancement task. As supported with the experimental results, the proposed method improves the quality of DDBTC decoded image.