BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation
This repository contain implementation of BiSeNet V2 in Tensorflow/Keras.
Instantiation of the Detail Branch and Semantic Branch. Each stage S contains one or more operations opr
(e.g., Conv2d, Stem, GE, CE). Each operation has a kernels size k, stride s and output channels c, repeated r times. The expansion
factor e is applied to expand the channel number of the operation. Here the channel ratio is ɑ = 1/4. The green colors mark
fewer channels of Semantic Branch in the corresponding stage of the Detail Branch. Notation: Conv2d means the convolutional
layer, followed by one batch normalization layer and relu activation function. Stem indicates the stem block. GE represents the
gather-and-expansion layer. CE is the context embedding block.
Dependies:
Tensorflow 2.0 or later
Licensed under MIT License