Distorted Representation Space Characterization Through Backpropagated Gradients
| Autor: | Gukyeong Kwon, Ghassan AlRegib, Mohit Prabhushankar, Dogancan Temel |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
FOS: Computer and information sciences
Standard test image Computer science Image quality business.industry Computer Vision and Pattern Recognition (cs.CV) Deep learning Image and Video Processing (eess.IV) Computer Science - Computer Vision and Pattern Recognition ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION 020206 networking & telecommunications Pattern recognition 02 engineering and technology Iterative reconstruction Electrical Engineering and Systems Science - Image and Video Processing Regularization (mathematics) Backpropagation FOS: Electrical engineering electronic engineering information engineering 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence Representation (mathematics) business |
| Zdroj: | ICIP |
| DOI: | 10.1109/icip.2019.8803228 |
| Popis: | In this paper, we utilize weight gradients from backpropagation to characterize the representation space learned by deep learning algorithms. We demonstrate the utility of such gradients in applications including perceptual image quality assessment and out-of-distribution classification. The applications are chosen to validate the effectiveness of gradients as features when the test image distribution is distorted from the train image distribution. In both applications, the proposed gradient based features outperform activation features. In image quality assessment, the proposed approach is compared with other state of the art approaches and is generally the top performing method on TID 2013 and MULTI-LIVE databases in terms of accuracy, consistency, linearity, and monotonic behavior. Finally, we analyze the effect of regularization on gradients using CURE-TSR dataset for out-of-distribution classification. 5 pages, 5 figures, 2 tables, ICIP 2019 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|