Add application files

app.py

@@ -0,0 +1,98 @@
+# ! pip install gradio
+
+import gradio as gr
+
+import tensorflow as tf 
+from tensorflow import keras
+from tensorflow.keras.models import Model, load_model
+
+import numpy as np
+
+import cv2
+
+from PIL import Image
+
+import matplotlib.pyplot as plt
+import matplotlib.patches as mpatches
+
+from pathlib import Path
+
+current_directory_path = Path(__file__).parent.resolve()
+object_detection_model_path = current_directory_path + "/" + "carla-image-segmentation-model.h5"
+lane_detection_model_path = current_directory_path + "/" + "lane-detection-for-carla-model.h5"
+
+label_map_object = {0: 'Unlabeled', 1: 'Building', 2: 'Fence', 3: 'Other', 
+                             4: 'Pedestrian', 5: 'Pole', 6: 'RoadLine', 7: 'Road', 8: 'SideWalk',
+                             9: 'Vegetation', 10: 'Vehicles', 11: 'Wall', 12: 'TrafficSign'}
+
+lane_label_map = {0: 'Unlabeled', 1: 'Left Lane', 2: 'Right Lane'}
+
+# Load the object detection model
+object_detection_model = load_model(object_detection_model_path)
+
+# Load the lane detection model
+lane_detection_model = load_model(lane_detection_model_path)
+
+
+def create_mask(object_detection_model, lane_detection_model, image):
+    # tensor = tf.convert_to_tensor(image, dtype=tf.float32)
+
+    image = tf.io.read_file(image.name)
+    image = tf.image.decode_png(image, channels=3)
+    image = tf.image.convert_image_dtype(image, tf.float32)
+    tensor = tf.image.resize(image, (256, 256), method='nearest')
+
+    # convert to tensor (specify 3 channels explicitly since png files contains additional alpha channel)
+    # set the dtypes to align with pytorch for comparison since it will use uint8 by default
+    # tensor = tf.io.decode_image(image_tensor, channels=3, dtype=tf.float32)
+
+    # resize tensor to 224 x 224
+    # tensor = tf.image.resize(tensor, [256, 256])
+
+    # add another dimension at the front to get NHWC shape
+    input_tensor = tf.expand_dims(tensor, axis=0)
+
+    # with mp_selfie.SelfieSegmentation(model_selection=0) as model:
+    # Create Masks for with Object Detection Model  
+    pred_masks_object_detect = object_detection_model.predict(input_tensor)
+    pred_masks_object_detect = tf.expand_dims(tf.argmax(pred_masks_object_detect, axis=-1), axis=-1)
+    pred_masks_object_detect = np.array(pred_masks_object_detect)
+
+    # Create Masks for with Lane Detection Model  
+    pred_masks_lane_detect = lane_detection_model.predict(input_tensor)
+    pred_masks_lane_detect = tf.expand_dims(tf.argmax(pred_masks_lane_detect, axis=-1), axis=-1)
+    pred_masks_lane_detect = np.array(pred_masks_lane_detect)
+    
+    return pred_masks_object_detect, pred_masks_lane_detect
+
+
+def segment_object(image): 
+    pred_masks_object_detect, pred_masks_lane_detect = create_mask(object_detection_model, lane_detection_model, image)
+
+    # image = cv2.resize(image, dsize=(256, 256), interpolation=cv2.INTER_CUBIC)
+
+    used_classes_object = np.unique(pred_masks_object_detect[0])
+    used_classes_lane = np.unique(pred_masks_lane_detect[0])
+
+    fig_object = plt.figure()
+    im = plt.imshow(tf.keras.preprocessing.image.array_to_img(pred_masks_object_detect[0]))            
+    patches_1 = [mpatches.Patch(color=im.cmap(im.norm(int(cls))), label="{}".format(label_map_object[int(cls)])) for cls in used_classes_object]
+    # put those patched as legend-handles into the legend
+    plt.legend(handles=patches_1, bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
+    plt.axis("off")   
+
+    fig_lane = plt.figure()
+    im = plt.imshow(tf.keras.preprocessing.image.array_to_img(pred_masks_lane_detect[0]))            
+    patches_1 = [mpatches.Patch(color=im.cmap(im.norm(int(cls))), label="{}".format(lane_label_map[int(cls)])) for cls in used_classes_lane]
+    # put those patched as legend-handles into the legend
+    plt.legend(handles=patches_1, bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
+    plt.axis("off")   
+
+    return fig_object
+
+
+webcam = gr.inputs.Image(shape=(800, 600), source="upload", type='file') #upload
+
+webapp = gr.interface.Interface(fn=segment_object, inputs=webcam, outputs="plot") #, live=False
+
+webapp.launch(debug=True)

carla-image-segmentation-model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b7f72d74610f5f8d39be28970b66a9aee8ccb8b50a768c57fa06803a62c4fbcc
+size 104177428

lane-detection-for-carla-model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9716b12103b745114ec39a54479eb2c38697925dfef06e88557354a08aa67bf1
+size 104173056