Run the following steps to implement semantic segmentation with DeepLabV3:
- Define the following function to load the pretrained DeepLab V3 model (frozen-inference graph) and run a forward pass with tensorflow to obtain the segmentation map:
def run_semantic_segmentation(image, model_path): input_tensor_name = 'ImageTensor:0' output_tensor_name = 'SemanticPredictions:0' input_size = 513 graph = tf.Graph() graph_def = None with gfile.FastGFile(model_path, 'rb') as f: graph_def = tf.GraphDef() graph_def.ParseFromString(f.read()) if graph_def is None: raise RuntimeError('Cannot find inference graph in tar \ archive.') with graph.as_default(): tf.import_graph_def(graph_def, name='') sess = tf.Session(graph=graph) ...
