Subventions et des contributions :

Subvention ou bourse octroyée s'appliquant à plus d'un exercice financier (2017-2018 à 2018-2019).

SPORTLOGiQ Inc. develops advanced sports analytics software to track the location and actions of playersx000D
appearing in sports game videos. This project focuses on the task of person re-identification in sports videos,x000D
where the same players should be recognized across multiple distributed cameras viewpoint, or across timex000D
within a single camera. In this context, the performance of person re-identification algorithms can decline duex000D
to variations in player capture conditions, background clutter, and occlusions. Computational complexity isx000D
another important consideration for real-time applications.x000D
This project aims to provide SPORTLOGiQ with the expertise needed to design accurate yet efficient systemsx000D
for real-time person re-identification. Given the state-of-the-art accuracy achieved with deep learningx000D
architectures on many challenging visual recognition problems, SPORTLOGiQ seeks to develop and evaluatex000D
deep convolutional neural networks (CNNs) for accurate person re-identification in sports videos. However,x000D
since these CNNs represent complex solutions for real-time applications, this project seeks to developx000D
specialized techniques for compression of CNNs, to reduce their time and memory complexity. In this project,x000D
the performance of several techniques proposed in the literature to increase the computational efficiency ofx000D
CNNs will be evaluated and compared for real-time person re-identification on sports game videos. Thesex000D
include advanced techniques for search reduction, features selection, and parameter pruning. In particular, thisx000D
project will focus mostly on promising new filter-level pruning techniques that can simultaneously acceleratex000D
and compress a CNN based on statistical information extracted from its layers. The techniques evaluated in thisx000D
project are of great interest to SPORTLOGiQ and to the computer vision and machine learning communities inx000D
general. The aim of this is research project is to solve one of the fundamental computer vision problems in thex000D
context of sports video analytics.