TeV makes available some of its datasets collected for research purposes.
SALSA Dataset - Multi-camera plus sociometric badges recording for multi-modal social scene analysis (60mins, 18 participants) with ground-truth and personality scores
CocktailParty Dataset - Multi-view dataset with calibration and ground-truth for the computer vision analysis of conversational groups
HALLWAY and LAB - Two datasets, also called Task-Decomposition. Through the sequences taken with four cameras in FBK Hall and TeV Lab respectively, people enter, walk around causing occlusions, sit down and exit randomly.
Image Enhancement Datasets and Software - Datasets of real-world color images, acquired by different devices, in indoor and outdoor environments, under different light conditions, including low-light and back-light
VENTURI Mountain Dataset - A collection of 12 outdoor sequences with camera pose ground-truth recorded with a smartphone
SmartWatch Gestures Dataset - Arm gesture recognition dataset recorded using the 3-axis accelerometer of a Sony SmartWatch™
DIPLODOC Road Stereo Sequence - Labeled road stereo sequence taken from an in-vehicle camera for evaluation of road detection algorithms