Connecting Look and Feel: Associating the visual and tactile properties of physical materials
This addresses the challenge for robots and machines to better understand physical material properties, though it is incremental as it builds on existing multimodal learning approaches.
The paper tackled the problem of associating visual and tactile properties of physical materials, specifically fabrics, by jointly training CNNs across color, depth, and tactile modalities, enabling prediction of feel from look and vice versa, with the system achieving improved performance over vision-only training when tested on visual inputs.
For machines to interact with the physical world, they must understand the physical properties of objects and materials they encounter. We use fabrics as an example of a deformable material with a rich set of mechanical properties. A thin flexible fabric, when draped, tends to look different from a heavy stiff fabric. It also feels different when touched. Using a collection of 118 fabric sample, we captured color and depth images of draped fabrics along with tactile data from a high resolution touch sensor. We then sought to associate the information from vision and touch by jointly training CNNs across the three modalities. Through the CNN, each input, regardless of the modality, generates an embedding vector that records the fabric's physical property. By comparing the embeddings, our system is able to look at a fabric image and predict how it will feel, and vice versa. We also show that a system jointly trained on vision and touch data can outperform a similar system trained only on visual data when tested purely with visual inputs.