no code implementations • 29 Nov 2022 • Heeseung Kwon, Francisco M. Castro, Manuel J. Marin-Jimenez, Nicolas Guil, Karteek Alahari
Vision Transformers (ViTs) have become a dominant paradigm for visual representation learning with self-attention operators.
no code implementations • 31 Aug 2020 • Zihao Mu, Francisco M. Castro, Manuel J. Marin-Jimenez, Nicolas Guil, Yan-ran Li, Shiqi Yu
In this paper, we propose iLGaCo, the first incremental learning approach of covariate factors for gait recognition, where the deep model can be updated with new information without re-training it from scratch by using the whole dataset.