摘要
We aim at advancing blind image quality assessment (BIQA), which predicts the human perception of image quality without any reference information. We develop a general and automated multitask learning scheme for BIQA to exploit auxiliary knowledge from other tasks, in a way that the model parameter sharing and the loss weighting are determined automatically. Specifically, we first describe all candidate label combinations (from multiple tasks) using a textual template, and compute the joint probability from the cosine similarities of the visual-textual embeddings. Predictions of each task can be inferred from the joint distribution, and optimized by carefully designed loss functions. Through comprehensive experiments on learning three tasks - BIQA, scene classification, and distortion type identification, we verify that the proposed BIQA method 1) benefits from the scene classification and distortion type identification tasks and outperforms the state-of-the-art on multiple IQA datasets, 2) is more robust in the group maximum differentiation competition, and 3) realigns the quality annotations from different IQA datasets more effectively. The source code is available at https://github.com/zwx8981/LIQE.
- Continual Learning for Blind Image Quality AssessmentZhang, Weixia; Li, Dingquan; Ma, Chao; Zhai, Guangtao; Yang, Xiaokang; Ma, Kede*IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(3): 2864-2878.被引 30
- A Multiscale Approach to Deep Blind Image Quality AssessmentLiu, Manni; Huang, Jiabin; Zeng, Delu*; Ding, Xinghao; Paisley, JohnIEEE Transactions on Image Processing, 2023, 32: 1656-1667.被引 4
- Toward a blind image quality evaluator in the wild by learning beyond human opinion scoresWang, Zhihua; Tang, Zhi-Ri; Zhang, Jianguo*; Fang, YumingPattern Recognition, 2023, 137: 109296.被引 1
