image_semantic_segmentation

README

Image Semantic Segmentation

The example uses DeepLab mobilenetv2_coco_voc_trainaug trained on PASCAL VOC 2012.

Refer to TensorFlow DeepLab Model Zoo and DeepLab: Deep Labelling for Semantic Image Segmentation for more details.

The input and output nodes of the model

Node Name	Input/Output	Shape	Data Description
ImageTensor	Input	[batch, height, width, 3]	RGB pixel values as uint8 in a square format (Width, Height)
SemanticPredictionss	Output	[batch, outeight, width]	Array of output segments

Usage

go run main.go -dir=<model folder> -jpg=<input.jpg> [-out=<output.jpg>] [-labels=<labels.txt>]

References

• [DeepLab Demo](https://github.com/tensorflow/models/blob/master/research/deeplab/deeplab_demo.ipynb
• TensorFlow visualization_utils
• An overview of semantic image segmentation