Uses differentiable NAS with temporal segments and pseudo-3D operators to discover a video action recognition network that outperforms hand-designed models on UCF101 with ~1% of the parameters when trained from scratch.
Deep residual learning for image recognition,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2019 2verdicts
UNVERDICTED 2representative citing papers
A quadruplet selection heuristic that pairs very hard negatives with relatively easy positives from matching hierarchical classes boosts embedding performance on fine-grained datasets.
citing papers explorer
-
Video Action Recognition Via Neural Architecture Searching
Uses differentiable NAS with temporal segments and pseudo-3D operators to discover a video action recognition network that outperforms hand-designed models on UCF101 with ~1% of the parameters when trained from scratch.
-
Quadruplet Selection Methods for Deep Embedding Learning
A quadruplet selection heuristic that pairs very hard negatives with relatively easy positives from matching hierarchical classes boosts embedding performance on fine-grained datasets.