Mahendran A, Vedaldi A. Understanding deep image representations by inverting them[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2015: 5188-5196.
1. Overview
In this paper, it proposes a general framework to invert representation
- invert a differentiable image representation using gradient descent
- first work to study the inverse of CNN image representation
1.1. Related Work
- DSIFT, HOG
- Exploring the representation capabilities of the HOG descriptor. show that it is possible to make any two images look nearly identical in SIFT space up to the injection of adversarial noise
- Intriguing properties of neural networks. demonstrated that adversarial example
1.2. Algorithm
- try to find the best x
- involve optimization at test time
1.3. Experiments
