PyTouch: A Machine Learning Library for Touch Processing
“A fundamental challenge in robotics is to process raw sensor measurements into high-level features that can be used for control and decision-making. For the sense of vision, the field of computer vision has been dedicated to study provide an algorithmic and programmatic method of understanding images and videos. In this field, open-source libraries such as PyTorch, CAFFE, and OpenCV have enabled the acceleration of research and collectively brought together commonly used techniques in each perspective domain by providing unified interfaces, algorithms, and platforms. PyTorch and CAFFE have enabled researchers to develop and scale neural networks by reducing the amount of ground work required for implementing algorithms such as backpropagation or convolution functions. Furthermore, OpenCV provided benefits such as a collection of commonly used, tested and optimized functions. These frameworks, libraries and platforms all have an underlying goal in enabling further developments in their respective fields through abstracting lower level processes into building blocks for further applications…”
Source: https://arxiv.org/pdf/2105.12791.pdf