ScaleNet

Introduced by Li et al. in Data-Driven Neuron Allocation for Scale Aggregation Networks

ScaleNet, or a Scale Aggregation Network, is a type of convolutional neural network which learns a neuron allocation for aggregating multi-scale information in different building blocks of a deep network. The most informative output neurons in each block are preserved while others are discarded, and thus neurons for multiple scales are competitively and adaptively allocated. The scale aggregation (SA) block concatenates feature maps at a wide range of scales. Feature maps for each scale are generated by a stack of downsampling, convolution and upsampling operations.

Source: Data-Driven Neuron Allocation for Scale Aggregation Networks

Read Paper See Code

Papers

Paper	Code	Results	Date	Stars

Tasks

Task	Papers	Share
Pose Estimation	2	22.22%
3D Reconstruction	1	11.11%
Geometric Matching	1	11.11%
Multi-Person Pose Estimation	1	11.11%
One-Shot Learning	1	11.11%
Semantic Segmentation	1	11.11%
Image Classification	1	11.11%
Object Detection	1	11.11%

Usage Over Time

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

Components

Component	Type	Add Remove
Bottleneck Residual Block	Skip Connection Blocks
Convolution	Convolutions
Dense Connections	Feedforward Networks
Global Average Pooling	Pooling Operations
Max Pooling	Pooling Operations
Scale Aggregation Block	Image Model Blocks
Softmax	Output Functions

Categories

Add Remove

Convolutional Neural Networks