Perceptual Video Coding using Deep Neural Network Based JND Model
Jong-Ho Kim, Dae Yeol Lee, Seyoon Jeong, Seung-Hyun Cho · 2020
We propose a perceptual video coding (PVC) method that uses the deep neural network (DNN) based just noticeable difference (JND) suppression model. The proposed JND suppression model's goal is to reduce the perceptual redundancy of the input video prior to the encoding process through a DNN, and further improve the compression efficiency while minimally affecting the perceptual quality.