SENet

Introduced by Hu et al. in Squeeze-and-Excitation Networks

A SENet is a convolutional neural network architecture that employs squeeze-and-excitation blocks to enable the network to perform dynamic channel-wise feature recalibration.

Source: Squeeze-and-Excitation Networks

Read Paper See Code

Papers

Paper	Code	Results	Date	Stars

Tasks

Task	Papers	Share
Image Classification	9	14.52%
Semantic Segmentation	6	9.68%
Object Detection	6	9.68%
Classification	4	6.45%
Instance Segmentation	4	6.45%
Clustering	2	3.23%
General Classification	2	3.23%
Emotion Recognition	2	3.23%
Few-Shot Learning	1	1.61%

Usage Over Time

This feature is experimental; we are continuously improving our matching algorithm.

Components

Component	Type	Add Remove
Convolution	Convolutions
Dense Connections	Feedforward Networks
Global Average Pooling	Pooling Operations
Kaiming Initialization	Initialization
Max Pooling	Pooling Operations
Softmax	Output Functions
Squeeze-and-Excitation Block	Image Model Blocks

Categories

Add Remove

Convolutional Neural Networks