ExoNet: Egocentric Images of Walking Environments

Egocentric vision is important for environment-adaptive control of humans and robots. Here we developed ExoNet, the largest open-source dataset of wearable camera images of real-world walking environments. The dataset contains over 5.6 million RGB images of indoor and outdoor environments, which were collected during summer, fall, and winter seasons. Over 923,000 images were human-annotated using a 12-class hierarchical labelling architecture. ExoNet serves as a communal platform to develop, optimize, and compare new deep learning models for egocentric visual perception, with applications in robotics and neuroscience. 

Reference:

Laschowski B, McNally W, Wong A, and McPhee J. (2022). Environment classification for robotic leg prostheses and exoskeletons using deep convolutional neural networks. Frontiers in Neurorobotics. DOI: 10.3389/fnbot.2021.730965.