Please click Details in the bottom right corner of each image for the details.

Using Monocular Vision and Human Body Priors for AUVs to Autonomously Approach Divers : Detect a scuba diver and approach to the diver using monocular vision and shoulder width information

ROW-SLAM (Under-Canopy Cornfield Semantic SLAM) : Detect corn stalks and localize them in a global coordinate frame

Semantically-Aware Strategies for Stereo-Visual Robotic Obstacle Avoidance : An obstacle avoidance module that combines visual instance segmentation with a depth map

Visual Diver Face Recognition : The first underwater diver face recognition system

LoCO : Low-Cost, Open-Source, Autonomous Underwater Vehicle (AUV)

TrashCAN dataset : Instance segmentation dataset

Trash VAE : Generating realistic trash data to overcome data scarcity

dMCL : Bathymetry-based Localization of Autonomous Underwater Robots

Detection of Marine Debris : Robotic Detection of Marine Litter Using Deep Visual Detection Models