ModelSqueeze-and-Excitation Network
SENet
A convolutional neural network architecture built on the SE (Squeeze-and-Excitation) module, which enhances feature representation through a channel attention mechanism. SENet won the ILSVRC 2017 classification challenge.
Timeline (last 90 days)
Oct 14
SENet won the ILSVRC 2017 championship thanks to the SE module.
Unverified50%
Aug 23
SENet引入了Squeeze-and-Excitation模块实现通道注意力机制
Unverified50%