Spaces:
Running
Running
Commit
•
864d3c8
1
Parent(s):
2a2a4bd
Upload 4 files
Browse files- ModelTransferLearning.py +99 -0
- PredictFace.py +42 -0
- config.py +6 -0
- trial1.py +243 -0
ModelTransferLearning.py
ADDED
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import numpy as np
|
2 |
+
import pandas as pd
|
3 |
+
import csv
|
4 |
+
import tensorflow as tf
|
5 |
+
from sklearn.model_selection import train_test_split
|
6 |
+
import cv2
|
7 |
+
from pathlib import Path
|
8 |
+
from tensorflow.keras.models import Sequential
|
9 |
+
from tensorflow.keras.layers import Dense, Flatten, Input
|
10 |
+
from tensorflow.keras.optimizers import Adam
|
11 |
+
from keras.applications import vgg16
|
12 |
+
|
13 |
+
def ModelFineTuning():
|
14 |
+
# Define the path to your dataset
|
15 |
+
data_dir = Path('Dataset')
|
16 |
+
image_size = (224, 224) # VGGFace model expects 224x224 images
|
17 |
+
|
18 |
+
# Initialize dictionaries
|
19 |
+
candidates_dict = {}
|
20 |
+
labels_dict = {}
|
21 |
+
|
22 |
+
# Get all class folder names
|
23 |
+
class_folders = [folder.name for folder in data_dir.iterdir() if folder.is_dir()]
|
24 |
+
total_classes = len(class_folders)
|
25 |
+
|
26 |
+
# Assign labels to each class
|
27 |
+
for idx, class_name in enumerate(class_folders):
|
28 |
+
candidates_dict[class_name] = list(data_dir.glob(f'{class_name}/*'))
|
29 |
+
labels_dict[class_name] = idx
|
30 |
+
|
31 |
+
df = pd.DataFrame(list(labels_dict.items()), columns=['Candidate Name', 'Label'])
|
32 |
+
df.to_csv("candidate_labels.csv", index=False)
|
33 |
+
|
34 |
+
# Print the results
|
35 |
+
print('Images Dictionary:')
|
36 |
+
print(candidates_dict)
|
37 |
+
print('\nLabels Dictionary:')
|
38 |
+
print(labels_dict)
|
39 |
+
|
40 |
+
X, y = [], []
|
41 |
+
if len(candidates_dict.items()) == 0:
|
42 |
+
return False
|
43 |
+
for candidate_name, faces in candidates_dict.items():
|
44 |
+
for image in faces:
|
45 |
+
img = cv2.imread(str(image))
|
46 |
+
resized_img = cv2.resize(img, image_size)
|
47 |
+
X.append(resized_img)
|
48 |
+
y.append(labels_dict[candidate_name])
|
49 |
+
|
50 |
+
print(len(X))
|
51 |
+
|
52 |
+
X = np.array(X)
|
53 |
+
y = np.array(y)
|
54 |
+
|
55 |
+
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)
|
56 |
+
|
57 |
+
X_train_scaled = X_train / 255.0
|
58 |
+
X_test_scaled = X_test / 255.0
|
59 |
+
|
60 |
+
# Convert labels to one-hot encoding
|
61 |
+
y_train = tf.keras.utils.to_categorical(y_train, num_classes=total_classes)
|
62 |
+
y_test = tf.keras.utils.to_categorical(y_test, num_classes=total_classes)
|
63 |
+
|
64 |
+
# Load the pre-trained VGGFace model
|
65 |
+
base_model = vgg16.VGG16(weights='imagenet', include_top=False, input_shape=(224, 224, 3))
|
66 |
+
|
67 |
+
# Ensure the base model layers are not trainable
|
68 |
+
for layer in base_model.layers:
|
69 |
+
layer.trainable = False
|
70 |
+
|
71 |
+
# Create a Sequential model and add layers
|
72 |
+
model = Sequential()
|
73 |
+
model.add(Input(shape=(224, 224, 3)))
|
74 |
+
model.add(base_model)
|
75 |
+
model.add(Flatten())
|
76 |
+
model.add(Dense(1024, activation='relu'))
|
77 |
+
model.add(Dense(512, activation='relu'))
|
78 |
+
model.add(Dense(total_classes, activation='softmax'))
|
79 |
+
|
80 |
+
# Compile the model
|
81 |
+
model.compile(optimizer=Adam(learning_rate=0.0001), loss='categorical_crossentropy', metrics=['accuracy'])
|
82 |
+
|
83 |
+
# Train the model
|
84 |
+
history = model.fit(
|
85 |
+
X_train_scaled, y_train,
|
86 |
+
validation_data=(X_test_scaled, y_test),
|
87 |
+
epochs=10, # Adjust the number of epochs based on your needs
|
88 |
+
batch_size=32
|
89 |
+
)
|
90 |
+
|
91 |
+
# Evaluate the model
|
92 |
+
loss, accuracy = model.evaluate(X_test_scaled, y_test)
|
93 |
+
print(f"Test accuracy: {accuracy * 100:.2f}%")
|
94 |
+
|
95 |
+
# Save the fine-tuned model
|
96 |
+
model.save('fine_tuned_VGG16_model.h5')
|
97 |
+
return True
|
98 |
+
|
99 |
+
# ModelFineTuning() # Uncomment this line to run the training
|
PredictFace.py
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import numpy as np
|
2 |
+
import pandas as pd
|
3 |
+
import tensorflow as tf
|
4 |
+
import cv2
|
5 |
+
from pathlib import Path
|
6 |
+
from datetime import datetime
|
7 |
+
from tensorflow.keras.models import load_model
|
8 |
+
|
9 |
+
|
10 |
+
# Function to preprocess the image
|
11 |
+
def preprocess_image(image_path, target_size=(224, 224)):
|
12 |
+
img = cv2.imread(image_path)
|
13 |
+
img = cv2.resize(img, target_size)
|
14 |
+
img = img / 255.0
|
15 |
+
img = np.expand_dims(img, axis=0)
|
16 |
+
return img
|
17 |
+
|
18 |
+
# Function to predict the label of the image
|
19 |
+
def predict_candidate(image_path):
|
20 |
+
|
21 |
+
model = load_model('fine_tuned_VGG16_model.h5')
|
22 |
+
|
23 |
+
df = pd.read_csv('candidate_labels.csv')
|
24 |
+
labels_dict = df.set_index('Label')['Candidate Name'].to_dict()
|
25 |
+
|
26 |
+
if image_path == '':
|
27 |
+
image_path = 'Dataset/test_img.jpg'
|
28 |
+
|
29 |
+
img = preprocess_image(image_path)
|
30 |
+
prediction = model.predict(img)
|
31 |
+
predicted_class = np.argmax(prediction, axis=1)[0]
|
32 |
+
predicted_label = labels_dict[predicted_class]
|
33 |
+
|
34 |
+
# Get the current timestamp in 12-hour format without seconds
|
35 |
+
timestamp = datetime.now().strftime('%Y-%m-%d %I:%M %p')
|
36 |
+
# Save to CSV
|
37 |
+
attendance_record = pd.DataFrame([[timestamp, predicted_label]], columns=['Timestamp', 'Student ID Number'])
|
38 |
+
attendance_record.to_csv('Attendance_Record.csv', mode='a', header=False, index=False)
|
39 |
+
|
40 |
+
print("Label is", predicted_label)
|
41 |
+
return predicted_label
|
42 |
+
|
config.py
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# config.py
|
2 |
+
import os
|
3 |
+
|
4 |
+
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
|
5 |
+
TF_OPS_PATH = os.path.join(BASE_DIR, 'path', 'to', 'tensorflow', 'python', 'ops')
|
6 |
+
|
trial1.py
ADDED
@@ -0,0 +1,243 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from flask import Flask, render_template, request, jsonify
|
2 |
+
import os
|
3 |
+
import base64
|
4 |
+
import numpy as np
|
5 |
+
import cv2
|
6 |
+
import tensorflow as tf
|
7 |
+
import csv
|
8 |
+
import config
|
9 |
+
from ModelTransferLearning import ModelFineTuning
|
10 |
+
from watchdog.observers import Observer
|
11 |
+
from watchdog.events import FileSystemEventHandler
|
12 |
+
import base64
|
13 |
+
from io import BytesIO
|
14 |
+
from PIL import Image
|
15 |
+
from PredictFace import preprocess_image, predict_candidate
|
16 |
+
|
17 |
+
app = Flask(__name__)
|
18 |
+
|
19 |
+
# Load the Caffe model
|
20 |
+
net = cv2.dnn.readNetFromCaffe('models/deploy.prototxt', 'models/res10_300x300_ssd_iter_140000.caffemodel')
|
21 |
+
|
22 |
+
###
|
23 |
+
def save_att_photos(images):
|
24 |
+
output_folder = 'AttendanceCapture'
|
25 |
+
if not os.path.exists(output_folder):
|
26 |
+
os.makedirs(output_folder)
|
27 |
+
print(f"Created directory: {output_folder}")
|
28 |
+
else:
|
29 |
+
print(f"Directory already exists: {output_folder}")
|
30 |
+
|
31 |
+
count = 0
|
32 |
+
for i, image_data in enumerate(images):
|
33 |
+
try:
|
34 |
+
print(f"Processing image {i + 1}")
|
35 |
+
image_data = base64.b64decode(image_data.split(',')[1])
|
36 |
+
image_np = np.frombuffer(image_data, np.uint8)
|
37 |
+
image = cv2.imdecode(image_np, cv2.IMREAD_COLOR)
|
38 |
+
if image is None:
|
39 |
+
print(f"Failed to decode image {i + 1}")
|
40 |
+
continue
|
41 |
+
|
42 |
+
(h, w) = image.shape[:2]
|
43 |
+
blob = cv2.dnn.blobFromImage(image, 1.0, (300, 300), [104, 117, 123], False, False)
|
44 |
+
net.setInput(blob)
|
45 |
+
detections = net.forward()
|
46 |
+
|
47 |
+
for j in range(detections.shape[2]):
|
48 |
+
confidence = detections[0, 0, j, 2]
|
49 |
+
if confidence > 0.7:
|
50 |
+
box = detections[0, 0, j, 3:7] * np.array([w, h, w, h])
|
51 |
+
(startX, startY, endX, endY) = box.astype("int")
|
52 |
+
face = image[startY:endY, startX:endX]
|
53 |
+
cv2.rectangle(image, (startX, startY), (endX, endY), (0, 255, 0), 2)
|
54 |
+
|
55 |
+
count += 1
|
56 |
+
face_filename = os.path.join(output_folder, f'face_{count}.jpg')
|
57 |
+
cv2.imwrite(face_filename, face)
|
58 |
+
print(f'Saved {face_filename}')
|
59 |
+
break
|
60 |
+
except Exception as e:
|
61 |
+
print(f"Error processing image {i + 1}: {e}")
|
62 |
+
continue
|
63 |
+
|
64 |
+
print(f"Total faces saved: {count}")
|
65 |
+
# save_to_csv(student_id, name)
|
66 |
+
return count >= 1
|
67 |
+
###
|
68 |
+
|
69 |
+
def save_photos(student_id, images):
|
70 |
+
output_folder = f'Dataset/{student_id}'
|
71 |
+
if not os.path.exists(output_folder):
|
72 |
+
os.makedirs(output_folder)
|
73 |
+
print(f"Created directory: {output_folder}")
|
74 |
+
else:
|
75 |
+
print(f"Directory already exists: {output_folder}")
|
76 |
+
|
77 |
+
count = 0
|
78 |
+
for i, image_data in enumerate(images):
|
79 |
+
try:
|
80 |
+
print(f"Processing image {i + 1}")
|
81 |
+
image_data = base64.b64decode(image_data.split(',')[1])
|
82 |
+
image_np = np.frombuffer(image_data, np.uint8)
|
83 |
+
image = cv2.imdecode(image_np, cv2.IMREAD_COLOR)
|
84 |
+
if image is None:
|
85 |
+
print(f"Failed to decode image {i + 1}")
|
86 |
+
continue
|
87 |
+
|
88 |
+
(h, w) = image.shape[:2]
|
89 |
+
blob = cv2.dnn.blobFromImage(image, 1.0, (300, 300), [104, 117, 123], False, False)
|
90 |
+
net.setInput(blob)
|
91 |
+
detections = net.forward()
|
92 |
+
|
93 |
+
for j in range(detections.shape[2]):
|
94 |
+
confidence = detections[0, 0, j, 2]
|
95 |
+
if confidence > 0.7:
|
96 |
+
box = detections[0, 0, j, 3:7] * np.array([w, h, w, h])
|
97 |
+
(startX, startY, endX, endY) = box.astype("int")
|
98 |
+
face = image[startY:endY, startX:endX]
|
99 |
+
cv2.rectangle(image, (startX, startY), (endX, endY), (0, 255, 0), 2)
|
100 |
+
|
101 |
+
count += 1
|
102 |
+
face_filename = os.path.join(output_folder, f'face_{count}.jpg')
|
103 |
+
cv2.imwrite(face_filename, face)
|
104 |
+
print(f'Saved {face_filename}')
|
105 |
+
break
|
106 |
+
except Exception as e:
|
107 |
+
print(f"Error processing image {i + 1}: {e}")
|
108 |
+
continue
|
109 |
+
|
110 |
+
print(f"Total faces saved: {count}")
|
111 |
+
# save_to_csv(student_id, name)
|
112 |
+
return count >= 20
|
113 |
+
|
114 |
+
|
115 |
+
@app.route('/start_capture', methods=['POST'])
|
116 |
+
def start_capture():
|
117 |
+
data = request.get_json()
|
118 |
+
student_id = data.get('student_id')
|
119 |
+
images = data.get('images')
|
120 |
+
|
121 |
+
if not student_id or not images:
|
122 |
+
return jsonify({'success': False, 'message': 'Invalid data'})
|
123 |
+
|
124 |
+
try:
|
125 |
+
if save_photos(student_id, images):
|
126 |
+
return jsonify({'success': True, 'message': 'Images saved successfully'})
|
127 |
+
else:
|
128 |
+
return jsonify({'success': False, 'message': 'Failed to save sufficient images'})
|
129 |
+
|
130 |
+
except Exception as e:
|
131 |
+
print(f'Error: {e}')
|
132 |
+
return jsonify({'success': False, 'message': 'Failed to save images'})
|
133 |
+
|
134 |
+
|
135 |
+
|
136 |
+
@app.route('/model_training', methods=['POST'])
|
137 |
+
def model_training():
|
138 |
+
success = ModelFineTuning()
|
139 |
+
if success:
|
140 |
+
return jsonify(success=True, message="Model training completed successfully.")
|
141 |
+
else:
|
142 |
+
return jsonify(success=False, message="Model training failed.")
|
143 |
+
|
144 |
+
|
145 |
+
###
|
146 |
+
@app.route('/take_photo', methods=['POST'])
|
147 |
+
def take_photo():
|
148 |
+
data = request.get_json()
|
149 |
+
# student_id = data.get('student_id')
|
150 |
+
images = data.get('images')
|
151 |
+
|
152 |
+
try:
|
153 |
+
if save_att_photos(images):
|
154 |
+
return jsonify({'success': True, 'message': 'Image saved successfully'})
|
155 |
+
else:
|
156 |
+
return jsonify({'success': False, 'message': 'Failed to save image'})
|
157 |
+
|
158 |
+
except Exception as e:
|
159 |
+
print(f'Error: {e}')
|
160 |
+
return jsonify({'success': False, 'message': 'Failed to save image'})
|
161 |
+
|
162 |
+
|
163 |
+
@app.route('/face_prediction', methods=['POST'])
|
164 |
+
def face_prediction():
|
165 |
+
image_path = 'AttendanceCapture/face_1.jpg'
|
166 |
+
predicted_label = predict_candidate(image_path)
|
167 |
+
# success = predict_candidate(image_path)
|
168 |
+
if predicted_label:
|
169 |
+
return jsonify(success=True, predicted_label = predicted_label)
|
170 |
+
else:
|
171 |
+
return jsonify(success=False, message="Attendance failed")
|
172 |
+
|
173 |
+
###
|
174 |
+
|
175 |
+
|
176 |
+
|
177 |
+
@app.route('/check_student_id', methods=['GET'])
|
178 |
+
def check_student_id():
|
179 |
+
student_id = request.args.get('student_id')
|
180 |
+
directory_exists = os.path.exists(f'Dataset/{student_id}')
|
181 |
+
return jsonify({"directory_exists": directory_exists})
|
182 |
+
|
183 |
+
|
184 |
+
|
185 |
+
@app.route("/register", methods=["POST", "GET"])
|
186 |
+
def register():
|
187 |
+
if request.method == "POST":
|
188 |
+
name = request.form["name"]
|
189 |
+
student_id = request.form["student_id"]
|
190 |
+
if os.path.exists(f'Dataset/{student_id}'):
|
191 |
+
return jsonify({"success": False, "message": "Student is already registered."})
|
192 |
+
else:
|
193 |
+
return jsonify({"success": True, "student_id": student_id})
|
194 |
+
return render_template("register.html")
|
195 |
+
|
196 |
+
|
197 |
+
|
198 |
+
@app.route('/capture')
|
199 |
+
def capture_photos():
|
200 |
+
student_id = request.args.get('student_id')
|
201 |
+
# name = request.args.get('name')
|
202 |
+
print(f"Capturing photos for student_id: {student_id}")
|
203 |
+
return render_template("capture_photos.html", student_id=student_id)
|
204 |
+
|
205 |
+
|
206 |
+
|
207 |
+
@app.route('/attendance')
|
208 |
+
def attendance():
|
209 |
+
return render_template("attendance.html")
|
210 |
+
|
211 |
+
|
212 |
+
|
213 |
+
@app.route('/')
|
214 |
+
def home():
|
215 |
+
return render_template("index.html")
|
216 |
+
|
217 |
+
|
218 |
+
# if __name__ == "__main__":
|
219 |
+
# app.run(debug=True)
|
220 |
+
|
221 |
+
|
222 |
+
class TrainingHandler(FileSystemEventHandler):
|
223 |
+
def on_modified(self, event):
|
224 |
+
print(f'File changed: {event.src_path}')
|
225 |
+
if 'custom_gradient.py' in event.src_path:
|
226 |
+
print("Restarting training...")
|
227 |
+
ModelFineTuning()
|
228 |
+
|
229 |
+
if __name__ == "__main__":
|
230 |
+
|
231 |
+
path_to_watch = os.environ.get('TF_OPS_PATH', 'C:\\Users\\Shruti Sundaram\\AppData\\Local\\Programs\\Python\\Python310\\Lib\\site-packages\\tensorflow\\python\\ops')
|
232 |
+
# path_to_watch = config.TF_OPS_PATH
|
233 |
+
event_handler = TrainingHandler()
|
234 |
+
observer = Observer()
|
235 |
+
observer.schedule(event_handler, path=path_to_watch, recursive=False)
|
236 |
+
observer.start()
|
237 |
+
|
238 |
+
try:
|
239 |
+
print("Starting initial training...")
|
240 |
+
app.run(debug=True)
|
241 |
+
except KeyboardInterrupt:
|
242 |
+
observer.stop()
|
243 |
+
observer.join()
|