no code implementations • 4 Mar 2020 • Shahine Bouabid, Vincent Delaitre
Mixup - a neural network regularization technique based on linear interpolation of labeled sample pairs - has stood out by its capacity to improve model's robustness and generalizability through a surprisingly simple formalism.
no code implementations • NeurIPS 2011 • Vincent Delaitre, Josef Sivic, Ivan Laptev
First, we replace the standard quantized local HOG/SIFT features with stronger discriminatively trained body part and object detectors.