TY - GEN
T1 - A Deep Blind Image Quality Assessment with Visual Importance Based Patch Score
AU - Lv, Zhengyi
AU - Wang, Xiaochuan
AU - Wang, Kai
AU - Liang, Xiaohui
N1 - Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
PY - 2019
Y1 - 2019
N2 - Convolutional neural networks (CNNs)-based no-reference image quality assessment (NR-IQA) suffers from insufficient training data. The conventional solution is splitting the training image into patches, assigning each patch the quality score, while the assignment of patch score is not consistent with the human visual system (HVS) well. To address the problem, we propose a patch quality assignment strategy, introducing the weighting map to describe the degree of visual importance of each distorted pixel, integrating the weighting map and the feature map to pool the quality score of each patch. With the patch quality, a CNNs-based NR-IQA model is trained. Experimental results demonstrate that proposed method, named as blind image quality metric with improved patch score (BIQIPS), improves the performance on most of the distortion types, especially on the types of local distortions, and achieves state-of-the-art prediction accuracy among the NR-IQA metrics.
AB - Convolutional neural networks (CNNs)-based no-reference image quality assessment (NR-IQA) suffers from insufficient training data. The conventional solution is splitting the training image into patches, assigning each patch the quality score, while the assignment of patch score is not consistent with the human visual system (HVS) well. To address the problem, we propose a patch quality assignment strategy, introducing the weighting map to describe the degree of visual importance of each distorted pixel, integrating the weighting map and the feature map to pool the quality score of each patch. With the patch quality, a CNNs-based NR-IQA model is trained. Experimental results demonstrate that proposed method, named as blind image quality metric with improved patch score (BIQIPS), improves the performance on most of the distortion types, especially on the types of local distortions, and achieves state-of-the-art prediction accuracy among the NR-IQA metrics.
KW - Convolutional neural networks
KW - Image quality assessment
KW - Patch score
KW - Visual importance
UR - https://www.scopus.com/pages/publications/85067349273
U2 - 10.1007/978-3-030-20890-5_10
DO - 10.1007/978-3-030-20890-5_10
M3 - 会议稿件
AN - SCOPUS:85067349273
SN - 9783030208899
T3 - Lecture Notes in Computer Science
SP - 147
EP - 162
BT - Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers
A2 - Jawahar, C.V.
A2 - Li, Hongdong
A2 - Mori, Greg
A2 - Schindler, Konrad
PB - Springer Verlag
T2 - 2018 Scene Understanding and Modelling Challenge, SUMO 2018, 2018Learning and Inference Methods for High-Performance Imaging, LIMHPI 2018, 2018 Attention/Intention Understanding, AIU 2018, 2018 Museum Exhibit Identification Challenge for Domain Adaptation and Few-Shot Learning, 2018 RGB-D—Sensing and Understanding via Combined Color and Depth, 2018 Dense 3D Reconstruction for Dynamic Scenes, 2018 AI Aesthetics in Art and Media, AIAM 2018, 3rd International Workshop on Robust Reading, IWRR 2018, 2018 Artificial Intelligence for Retinal Image Analysis, AIRIA 2018, 2018 Combining Vision and Language, 1st International Workshop on Advanced Machine Vision for Real-Life and Industrially Relevant Applications, AMV 2018
Y2 - 2 December 2018 through 6 December 2018
ER -