Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386076169> ?p ?o ?g. }
- W4386076169 abstract "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." @default.
- W4386076169 created "2023-08-23" @default.
- W4386076169 creator A5009811665 @default.
- W4386076169 creator A5019708391 @default.
- W4386076169 creator A5020029652 @default.
- W4386076169 creator A5064168853 @default.
- W4386076169 creator A5070942166 @default.
- W4386076169 date "2023-06-01" @default.
- W4386076169 modified "2023-10-16" @default.
- W4386076169 title "Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspective" @default.
- W4386076169 cites W1977725648 @default.
- W4386076169 cites W1979451680 @default.
- W4386076169 cites W1982471090 @default.
- W4386076169 cites W1997188340 @default.
- W4386076169 cites W2033442452 @default.
- W4386076169 cites W2051596736 @default.
- W4386076169 cites W2085518012 @default.
- W4386076169 cites W2102166818 @default.
- W4386076169 cites W2108598243 @default.
- W4386076169 cites W2129644086 @default.
- W4386076169 cites W2133665775 @default.
- W4386076169 cites W2138790992 @default.
- W4386076169 cites W2156935079 @default.
- W4386076169 cites W2161907179 @default.
- W4386076169 cites W2473697052 @default.
- W4386076169 cites W2556068545 @default.
- W4386076169 cites W2768340063 @default.
- W4386076169 cites W2897228451 @default.
- W4386076169 cites W2905544033 @default.
- W4386076169 cites W2906729185 @default.
- W4386076169 cites W2953590133 @default.
- W4386076169 cites W2962784628 @default.
- W4386076169 cites W2963430933 @default.
- W4386076169 cites W2963498646 @default.
- W4386076169 cites W2963877604 @default.
- W4386076169 cites W2963918210 @default.
- W4386076169 cites W2963975576 @default.
- W4386076169 cites W2966182616 @default.
- W4386076169 cites W2970470466 @default.
- W4386076169 cites W2970763616 @default.
- W4386076169 cites W2976809040 @default.
- W4386076169 cites W3035712445 @default.
- W4386076169 cites W3035719652 @default.
- W4386076169 cites W3036239693 @default.
- W4386076169 cites W3091249416 @default.
- W4386076169 cites W3100404621 @default.
- W4386076169 cites W3100498948 @default.
- W4386076169 cites W3104772192 @default.
- W4386076169 cites W3135479537 @default.
- W4386076169 cites W3149775056 @default.
- W4386076169 cites W3159198982 @default.
- W4386076169 cites W3174692951 @default.
- W4386076169 cites W3198377975 @default.
- W4386076169 cites W3207397767 @default.
- W4386076169 cites W3210514413 @default.
- W4386076169 cites W4214745154 @default.
- W4386076169 cites W4224267514 @default.
- W4386076169 cites W4286611322 @default.
- W4386076169 cites W4312354469 @default.
- W4386076169 cites W4312956471 @default.
- W4386076169 cites W4312980231 @default.
- W4386076169 doi "https://doi.org/10.1109/cvpr52729.2023.01352" @default.
- W4386076169 hasPublicationYear "2023" @default.
- W4386076169 type Work @default.
- W4386076169 citedByCount "0" @default.
- W4386076169 crossrefType "proceedings-article" @default.
- W4386076169 hasAuthorship W4386076169A5009811665 @default.
- W4386076169 hasAuthorship W4386076169A5019708391 @default.
- W4386076169 hasAuthorship W4386076169A5020029652 @default.
- W4386076169 hasAuthorship W4386076169A5064168853 @default.
- W4386076169 hasAuthorship W4386076169A5070942166 @default.
- W4386076169 hasConcept C111472728 @default.
- W4386076169 hasConcept C111919701 @default.
- W4386076169 hasConcept C115961682 @default.
- W4386076169 hasConcept C116834253 @default.
- W4386076169 hasConcept C119857082 @default.
- W4386076169 hasConcept C126780896 @default.
- W4386076169 hasConcept C126838900 @default.
- W4386076169 hasConcept C12713177 @default.
- W4386076169 hasConcept C138885662 @default.
- W4386076169 hasConcept C153180895 @default.
- W4386076169 hasConcept C154945302 @default.
- W4386076169 hasConcept C162324750 @default.
- W4386076169 hasConcept C165696696 @default.
- W4386076169 hasConcept C183115368 @default.
- W4386076169 hasConcept C187736073 @default.
- W4386076169 hasConcept C194257627 @default.
- W4386076169 hasConcept C2776257435 @default.
- W4386076169 hasConcept C2779530757 @default.
- W4386076169 hasConcept C2780451532 @default.
- W4386076169 hasConcept C28006648 @default.
- W4386076169 hasConcept C31258907 @default.
- W4386076169 hasConcept C38652104 @default.
- W4386076169 hasConcept C41008148 @default.
- W4386076169 hasConcept C43126263 @default.
- W4386076169 hasConcept C55020928 @default.
- W4386076169 hasConcept C59822182 @default.
- W4386076169 hasConcept C71924100 @default.
- W4386076169 hasConcept C86803240 @default.
- W4386076169 hasConceptScore W4386076169C111472728 @default.