Zobrazeno 1 - 10
of 159
pro vyhledávání: '"Tomasi, Carlo"'
Although binary classification is a well-studied problem, training reliable classifiers under severe class imbalance remains a challenge. Recent techniques mitigate the ill effects of imbalance on training by modifying the loss functions or optimizat
Externí odkaz:
http://arxiv.org/abs/2410.03588
Although binary classification is a well-studied problem in computer vision, training reliable classifiers under severe class imbalance remains a challenging problem. Recent work has proposed techniques that mitigate the effects of training under imb
Externí odkaz:
http://arxiv.org/abs/2402.05400
Autor:
Yuan, Shuai, Tomasi, Carlo
Both optical flow and stereo disparities are image matches and can therefore benefit from joint training. Depth and 3D motion provide geometric rather than photometric information and can further improve optical flow. Accordingly, we design a first n
Externí odkaz:
http://arxiv.org/abs/2310.04712
Unsupervised optical flow estimation is especially hard near occlusions and motion boundaries and in low-texture regions. We show that additional information such as semantics and domain knowledge can help better constrain this problem. We introduce
Externí odkaz:
http://arxiv.org/abs/2303.06209
Unsupervised optical flow estimators based on deep learning have attracted increasing attention due to the cost and difficulty of annotating for ground truth. Although performance measured by average End-Point Error (EPE) has improved over the years,
Externí odkaz:
http://arxiv.org/abs/2208.02305
We propose TAIN (Transformers and Attention for video INterpolation), a residual neural network for video interpolation, which aims to interpolate an intermediate frame given two consecutive image frames around it. We first present a novel vision tra
Externí odkaz:
http://arxiv.org/abs/2207.04132
Supervised training of optical flow predictors generally yields better accuracy than unsupervised training. However, the improved performance comes at an often high annotation cost. Semi-supervised training trades off accuracy against annotation cost
Externí odkaz:
http://arxiv.org/abs/2203.05053
Publikováno v:
The British Machine Vision Conference (BMVC), 2021
We propose MONet, a convolutional neural network that jointly detects motion boundaries (MBs) and occlusion regions (Occs) in video both forward and backward in time. Detection is difficult because optical flow is discontinuous along MBs and undefine
Externí odkaz:
http://arxiv.org/abs/2111.01261
Autor:
Yu, Shuzhi, Tomasi, Carlo
Residual Neural Networks (ResNets) achieve state-of-the-art performance in many computer vision problems. Compared to plain networks without residual connections (PlnNets), ResNets train faster, generalize better, and suffer less from the so-called d
Externí odkaz:
http://arxiv.org/abs/1905.10944
Autor:
Ristani, Ergys, Tomasi, Carlo
Multi-Target Multi-Camera Tracking (MTMCT) tracks many people through video taken from several cameras. Person Re-Identification (Re-ID) retrieves from a gallery images of people similar to a person query image. We learn good features for both MTMCT
Externí odkaz:
http://arxiv.org/abs/1803.10859