Blog

Blog

Learning to Generate Chairs with Convolutional Neural Networks

For a project during my masters degree, we implemented the paper Learning to Generate Chairs by Dosovitskiy et. al. in PyTorch. The only implementation available was the author’s in Caffe with Lua, so we set off to create a clean open source implementation of the work. This work presents an early differentiable renderer using convolutional…

Paper reading: SDM-NET & New tools for 3D deep learning

1.This week I found a paper, SDM-NET: Deep Generative Network for Structured Deformable Mesh. Their main idea is to decompose a 3D object into parts. And each part can be deformed from a homeomorphic simple object. Their network can also learn the support and symmetry information from the input. The part structures and geometries are…

Tested a 15m Walkway

This week I setup a 15-meter walkway in the lab for the obstacle experiment in walking paradigm. I chose to use Vive pro with a wireless adapter to accomplish a wireless VR experience. Compared with Oculus Quest, the wireless Vive pro is much heavier, and it is still constrained by the coverage of Lighthouses and…

A parallel world

This last week or so I’ve been building a parallel world. I measured all of the sizes in a section of our lab — walls, floors, tables, and so forth. Then I built a computer graphic copy of that world, matching the real thing down to the centimeter. I can walk around in my parallel…

Virtual Obstacles in Gait

A few weeks earlier, I talked about two papers I read regarding gait balance when fixed or suddenly appeared obstacles are involved. The papers are Age Effects on Strategies Used to Avoid Obstacles, and Anxiety-mediated gait adaptations reduce errors of obstacle negotiation among younger and older adults: Implications for fall risk. We have been looking for protocols and…