Videos For: One-shot recognition of any material anywhere using contrastive learning with physics-based renderin

Demonstration Video: Image Recognition of Materials States And Types From Single Example Identifying the state of the material in the main video by matching it to a few reference images (top). Each image contains a specific material state or type. The best match image is marked green. Based on the net for identifying similarity between any material state and type. The code for the net and the trained model used for this is available at: https://github.com/sagieppel/Contrastive-learning-for-one-shot-materials-and-textures-similarity-recognition-from-images Paper : One-shot recognition of any material anywhere using contrastive learning with physics-based rendering. https://arxiv.org/abs/2212.00648