Research Pages

A Benchmark and Simulator for UAV Tracking (Simulator)

UAV Simulator: You can download the complete project binaries including a free demo world here.

Activity Net

In spite of many dataset efforts for human action recognition, current computer vision algorithms are still severely limited in terms of the variability and complexity of the actions that they can recognize. This is in part due to the simplicity of current benchmarks, which mostly focus on simple actions and movements occurring on manually trimmed videos. We introduce ActivityNet, a new large-scale video benchmark for human activity understanding.

IVUL Publications

2024: GES: Generalized Exponential Splatting for Efficient Radiance Field Rendering, Abdullah Hamdi, Luke Melas-Kyriazi, Jinjie Mai, Guocheng Qian, Ruoshi Liu, Carl Vondrick, Bernard Ghanem, Andrea Vedaldi, Conference on Computer Vision and Pattern Recognition (CVPR 2024), ( Link) Dr2Net: Dynamic Reversible Dual-Residual Networks for Memory-Efficient Finetuning, Chen Zhao, Shuming Liu, Karttikeya Mangalam, Guocheng Qian, Fatimah Zohra, Abdulmohsen Alghannam, Jitendra Malik, Bernard Ghanem, Conference on Computer Vision and Pattern Recognition (CVPR 2024) SPAD: Spatially Aware Multiview

IVUL Research Group Collaborators

University of Amsterdam The University of Amsterdam (abbreviated as UvA, Dutch: Universiteit van Amsterdam) is a prestigious public university located in Amsterdam, Netherlands. Established in 1632 by municipal authorities and later renamed for the city of Amsterdam, the University of Amsterdam is the third-oldest university in the Netherlands. It is the largest university in the Netherlands by enrollment. Collaborators: Prof. Cees Snoek UC Merced UC Merced opened Sept. 5, 2005, as the newest campus in the University of California system and the first American research university of the

TrackingNet

Despite the numerous developments in object tracking, further improvement of current tracking algorithms is limited by small and mostly saturated datasets. As a matter of fact, data-hungry trackers based on deep-learning currently rely on object detection datasets due to the scarcity of dedicated large-scale tracking datasets. In this work, we present TrackingNet, the first large-scale dataset, and benchmark for object tracking in the wild. We provide more than 30K videos with more than 14 million dense bounding box annotations.