DenseNet

Introduced by Huang et al. in Densely Connected Convolutional Networks

A DenseNet is a type of convolutional neural network that utilises dense connections between layers, through Dense Blocks, where we connect all layers (with matching feature-map sizes) directly with each other. To preserve the feed-forward nature, each layer obtains additional inputs from all preceding layers and passes on its own feature-maps to all subsequent layers.

Source: Densely Connected Convolutional Networks

Read Paper See Code

Papers

Paper	Code	Results	Date	Stars

Tasks

Task	Papers	Share
Image Classification	59	13.88%
General Classification	58	13.65%
Semantic Segmentation	17	4.00%
Object Detection	14	3.29%
Quantization	7	1.65%
Computed Tomography (CT)	6	1.41%
Out-of-Distribution Detection	5	1.18%
Domain Adaptation	5	1.18%
reinforcement-learning	4	0.94%

Usage Over Time

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

Components

Component	Type	Add Remove
1x1 Convolution	Convolutions
Average Pooling	Pooling Operations
Convolution	Convolutions
Dense Block	Image Model Blocks
Dense Connections	Feedforward Networks
Dropout	Regularization
Global Average Pooling	Pooling Operations
Kaiming Initialization	Initialization
Max Pooling	Pooling Operations
Softmax	Output Functions

Categories

Add Remove

Convolutional Neural Networks