Distorted Representation Space Characterization Through Backpropagated Gradients
arXiv:1908.09998
Abstract
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.
arXiv comment5 pages, 5 figures, 2 tables, ICIP 2019
References in corpus (7)
- Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks
- Video Anomaly Detection and Localization via Gaussian Mixture Fully Convolutional Variational Autoencoder
- Perceptual Image Quality Assessment through Spectral Analysis of Error Representations
- CSV: Image Quality Assessment Based on Color, Structure, and Visual System
- PerSIM: Multi-resolution Image Quality Assessment in the Perceptually Uniform Color Domain
- Traffic Signs in the Wild: Highlights from the IEEE Video and Image Processing Cup 2017 Student Competition [SP Competitions]
- Challenging Environments for Traffic Sign Detection: Reliability Assessment under Inclement Conditions