Partial List of Implemented Layers

クラス
class	cv::dnn::BlankLayer
class	cv::dnn::ConstLayer
class	cv::dnn::LSTMLayer
	LSTM再帰層[【詳解】（英語］
class	cv::dnn::RNNLayer
	古典的な再帰層[【詳解】（英語］
class	cv::dnn::BaseConvolutionLayer
class	cv::dnn::ConvolutionLayer
class	cv::dnn::DeconvolutionLayer
class	cv::dnn::LRNLayer
class	cv::dnn::PoolingLayer
class	cv::dnn::SoftmaxLayer
class	cv::dnn::InnerProductLayer
class	cv::dnn::MVNLayer
class	cv::dnn::ReshapeLayer
class	cv::dnn::FlattenLayer
class	cv::dnn::ConcatLayer
class	cv::dnn::SplitLayer
class	cv::dnn::SliceLayer
class	cv::dnn::PermuteLayer
class	cv::dnn::ShuffleChannelLayer
class	cv::dnn::PaddingLayer
	特定の軸に対する追加の値を追加します。[【詳解】（英語］
class	cv::dnn::ActivationLayer
class	cv::dnn::ReLULayer
class	cv::dnn::ReLU6Layer
class	cv::dnn::ChannelsPReLULayer
class	cv::dnn::ELULayer
class	cv::dnn::TanHLayer
class	cv::dnn::SwishLayer
class	cv::dnn::MishLayer
class	cv::dnn::SigmoidLayer
class	cv::dnn::BNLLLayer
class	cv::dnn::AbsLayer
class	cv::dnn::PowerLayer
class	cv::dnn::ExpLayer
class	cv::dnn::CropLayer
class	cv::dnn::EltwiseLayer
	入力に対する要素ごとの操作[【詳解】（英語］
class	cv::dnn::BatchNormLayer
class	cv::dnn::MaxUnpoolLayer
class	cv::dnn::ScaleLayer
class	cv::dnn::ShiftLayer
class	cv::dnn::DataAugmentationLayer
class	cv::dnn::CorrelationLayer
class	cv::dnn::AccumLayer
class	cv::dnn::FlowWarpLayer
class	cv::dnn::PriorBoxLayer
class	cv::dnn::ReorgLayer
class	cv::dnn::RegionLayer
class	cv::dnn::DetectionOutputLayer
	検出出力層。[【詳解】（英語］
class	cv::dnn::NormalizeBBoxLayer
	- Normalization Layer (正規化層)[【詳解】（英語］
class	cv::dnn::ResizeLayer
	入力された4次元blobのサイズを、最近接法またはバイリニア法で変更します。[【詳解】（英語］
class	cv::dnn::InterpLayer
	のバイリニアリサイズレイヤーhttps://github.com/cdmh/deeplab-public-ver2 [【詳解】（英語］
class	cv::dnn::ProposalLayer
class	cv::dnn::CropAndResizeLayer

詳解

This subsection of dnn module contains information about built-in layers and their descriptions.

Classes listed here, in fact, provides C++ API for creating instances of built-in layers. In addition to this way of layers instantiation, there is a more common factory API (see Utilities for New Layers Registration), it allows to create layers dynamically (by name) and register new ones. You can use both API, but factory API is less convenient for native C++ programming and basically designed for use inside importers (see readNetFromCaffe(), readNetFromTorch(), readNetFromTensorflow()).

Built-in layers partially reproduce functionality of corresponding Caffe and Torch7 layers. In particular, the following layers and Caffe importer were tested to reproduce Caffe functionality:

• Convolution
• Deconvolution
• Pooling
• InnerProduct
• TanH, ReLU, Sigmoid, BNLL, Power, AbsVal
• Softmax
• Reshape, Flatten, Slice, Split
• LRN
• MVN
• Dropout (since it does nothing on forward pass -))