Upload 14 files

.idea/.gitignore

@@ -0,0 +1,3 @@
+# Default ignored files
+/shelf/
+/workspace.xml

.idea/inspectionProfiles/profiles_settings.xml

@@ -0,0 +1,6 @@
+<component name="InspectionProjectProfileManager">
+  <settings>
+    <option name="USE_PROJECT_PROFILE" value="false" />
+    <version value="1.0" />
+  </settings>
+</component>

.idea/misc.xml

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.9 (traffic_signs_classification)" project-jdk-type="Python SDK" />
+</project>

.idea/modules.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/traffic_signs_classification.iml" filepath="$PROJECT_DIR$/.idea/traffic_signs_classification.iml" />
+    </modules>
+  </component>
+</project>

.idea/traffic_signs_classification.iml

@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="PYTHON_MODULE" version="4">
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$">
+      <excludeFolder url="file://$MODULE_DIR$/venv" />
+    </content>
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>

README.md

@@ -1,12 +1,15 @@
----
-title: Trafficsignclassification
-emoji: 🔥
-colorFrom: indigo
-colorTo: blue
-sdk: streamlit
-sdk_version: 1.33.0
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# traffic_signs_classification
+
+The purpose of this project is to train an implementation of the LeNet-5 Convolutional Neural Network for the 
+classification of traffic signs provided by the user. The model was trained and tested on about 40.000 real photos of 43 
+types of german traffic signs. Data was taken from The German Traffic Sign Recognition Benchmark (GTSRB): https://benchmark.ini.rub.de/gtsrb_dataset.html
+
+Notebook with built and trained model: **traffic_signs_classification_lenet5.ipynb**
+
+Folder with saved model: **keras_model**
+
+Streamlit application for testing the model: **streamlit_app.py**
+
+Article: https://towardsdatascience.com/classification-of-traffic-signs-with-lenet-5-cnn-cb861289bd62
+
+Link to Streamlit app: https://share.streamlit.io/andriigoz/traffic_signs_classification

keras_model/keras_metadata.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53a3f046cf6060be29b41e6b77c656851ed5d3b5ce93e6672f4f238592b1364b
+size 19323

keras_model/saved_model.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:461ab68ffe048c5c8157e7dd7c9af66ccc02bb60024bfe01d6f77ee631760df5
+size 169181

requirements.txt

@@ -0,0 +1,7 @@
+keras==2.7.0
+numpy==1.21.0
+opencv-python-headless
+Pillow==9.0.0
+streamlit==1.4.0
+tensorflow==2.7.0
+click==8

streamlit_app.py

@@ -0,0 +1,121 @@
+import numpy as np
+import streamlit as st
+import cv2 as cv
+from PIL import Image
+from keras.models import load_model
+
+
+# Label traffic signs
+labels_dict = {
+    0: '274. Speed limit (20km/h)',
+    1: '274. Speed limit (30km/h)',
+    2: '274. Speed limit (50km/h)',
+    3: '274. Speed limit (60km/h)',
+    4: '274. Speed limit (70km/h)',
+    5: '274. Speed limit (80km/h)',
+    6: '278. End of speed limit (80km/h)',
+    7: '274. Speed limit (100km/h)',
+    8: '274. Speed limit (120km/h)',
+    9: '276. No passing',
+    10: '277. No passing vehicle over 3.5 tons',
+    11: '301. Right-of-way at intersection',
+    12: '306. Priority road',
+    13: '205. Yield',
+    14: '206. Stop',
+    15: '250. No vehicles',
+    16: '253. Vehicle > 3.5 tons prohibited',
+    17: '267. No entry',
+    18: '101. General caution',
+    19: '104. Dangerous curve left',
+    20: '103. Dangerous curve right',
+    21: '105-10. Double curve',
+    22: '112. Bumpy road',
+    23: '114. Slippery road',
+    24: '121. Road narrows on the right',
+    25: '123. Road work',
+    26: '131. Traffic signals',
+    27: '133. Pedestrians',
+    28: '136. Children crossing',
+    29: '138. Bicycles crossing',
+    30: '101-51. Beware of ice/snow',
+    31: '142. Wild animals crossing',
+    32: '282. End speed + passing limits',
+    33: '209-20. Turn right ahead',
+    34: '209-10. Turn left ahead',
+    35: '209-30. Ahead only',
+    36: '214-20. Go straight or right',
+    37: '214-10. Go straight or left',
+    38: '222-20. Keep right',
+    39: '222-10. Keep left',
+    40: '215. Roundabout mandatory',
+    41: '280. End of no passing',
+    42: '281. End no passing vehicle > 3.5 tons'
+}
+
+
+@st.cache
+def sign_predict(image):
+    model = load_model('keras_model\saved_model.pb')
+    image = np.array(image, dtype=np.float32)
+    image = image/255
+    image = np.reshape(image, (1, 32, 32))
+    x = image.astype(np.float32)
+    prediction = model.predict(x)
+    prediction_max = np.argmax(prediction)
+    prediction_label = labels_dict[prediction_max]
+    confidence = np.max(prediction)
+    return prediction_label, confidence
+
+
+def main():
+    # Set page config and markdowns
+    st.set_page_config(page_title='Traffic Signs Classifier', page_icon=':car:')
+    st.title('Traffic Signs Classifier')
+    st.markdown("""
+        This application classifies traffic signs. Upload any photo of a traffic sign 
+        and receive its name out of 43 present classes. For getting the correct prediction, 
+        try to upload a square picture containing only the sign.
+        """)
+    with st.expander("See list of classes"):
+        st.write(list(labels_dict.values()))
+    st.image('./app_images/road_sign.jpg', use_column_width=True)
+    image_usr = st.file_uploader('Upload a photo of traffic sign here', type=['jpg', 'jpeg', 'png'])
+
+    if image_usr is not None:
+        col1, col2 = st.columns(2)
+        col1.markdown('#### Your picture')
+        col2.markdown('#### Your picture 32x32 gray')
+        image = Image.open(image_usr)
+        with col1:
+            st.image(image, use_column_width=True)
+
+        image_np = np.array(image.convert('RGB'))
+        image_col = cv.cvtColor(image_np, 1)
+        image_gray = cv.cvtColor(image_col, cv.COLOR_BGR2GRAY)
+        image_32 = cv.resize(image_gray, (32, 32))
+        with col2:
+            st.image(image_32, use_column_width=True)
+
+        # Make prediction
+        prediction_label, confidence = sign_predict(image_32)
+
+        st.write('##### Prediction:', prediction_label)
+        st.write('##### Confidence:', str(confidence))
+        st.markdown('***')
+
+    # Markdowns
+    st.subheader('About this app')
+    st.markdown("""
+    The app uses an implementation of LeNet-5 Convolutional Neural Network. 
+    The model was trained and tested on about 40.000 real photos of 43 types of german traffic signs.
+    
+    Data was taken from The German Traffic Sign Recognition Benchmark (GTSRB):
+    https://benchmark.ini.rub.de/gtsrb_dataset.html
+    
+    
+    Author: Prajwal Thawari
+    """)
+
+
+if __name__ == '__main__':
+    main()