Physics-based Feature Extraction and Image Manipulation via Autoencoders
semanticscholar(2017)
摘要
We experiment with the extraction of physics-based features by utilizing synthesized data as ground truth, and further utilize these extracted features to perform image space manipulations. Inspired by recent literature, we model our network as an adversarial autoencoder, and train our encoder to extract physical properties of the image scene.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要