datasets

Labeled Fisheries Underwater Dataset

This dataset consists of more than 20,000 labeled images with four different underwater species – rock fish, starfish, sponge, and corals. It is built via an iterative labeling process utilizing a human-in-the-loop deep learning framework as described in our upcoming paper, “An Iterative Labeling Method for Annotating Imagery for Fisheries Stock Assessment.”

Fill out this form to access our labeled fisheries underwater dataset.

LIDAR Bicycle Datasets​

Dataset to support the thesis “Progress Towards LiDAR Based Bicycle Detection in Urban Environments.”

Download the rosbag data for this code here.