File size: 14,630 Bytes
85f4388
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
from flask import Flask, request, jsonify, render_template, url_for
from flask_cors import CORS
import torch
import torch.nn as nn
from torchvision import models, transforms
from PIL import Image
from huggingface_hub import hf_hub_download
import os
from mtcnn import MTCNN  
import cv2
from flask_bcrypt import generate_password_hash, check_password_hash
from pymongo import MongoClient
import numpy as np
from werkzeug.security import generate_password_hash, check_password_hash
from werkzeug.utils import secure_filename
import logging
import matplotlib.pyplot as plt
import seaborn as sns
from transformers import AutoImageProcessor, AutoModelForImageClassification  # New imports

# Setup logging
logging.basicConfig(level=logging.INFO)

app = Flask(__name__, template_folder="templates", static_folder="static")
CORS(app)

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
UPLOAD_FOLDER = "static/uploads"
os.makedirs(UPLOAD_FOLDER, exist_ok=True)

# ------------------- Model Loading Functions -------------------

def load_model_from_hf(repo_id, filename, num_classes):
    model_path = hf_hub_download(repo_id=repo_id, filename=filename)
    model = models.convnext_tiny(weights=None)
    in_features = model.classifier[2].in_features
    model.classifier[2] = nn.Linear(in_features, num_classes)
    model.load_state_dict(torch.load(model_path, map_location=device))
    model.to(device)
    model.eval()
    return model

# Load the existing deepfake/cheapfake models
deepfake_model = load_model_from_hf("faryalnimra/DFDC-detection-model", "DFDC.pth", 2)  
cheapfake_model = load_model_from_hf("faryalnimra/ORIG-TAMP", "ORIG-TAMP.pth", 1)

# ------------------- New Real/Fake Detector Model -------------------
# This model determines if the uploaded image is real (label 1) or fake (label 0)
model_name = "prithivMLmods/Deep-Fake-Detector-Model"
processor = AutoImageProcessor.from_pretrained(model_name, use_fast=False)
realfake_detector = AutoModelForImageClassification.from_pretrained(model_name)
realfake_detector.to(device)
realfake_detector.eval()

# ------------------- Image Preprocessing -------------------

transform = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])

# ------------------- Face Detector -------------------

face_detector = MTCNN()

def detect_face(image_path):
    image = cv2.imread(image_path)
    image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
    faces = face_detector.detect_faces(image_rgb)
    face_count = sum(1 for face in faces if face.get("confidence", 0) > 0.90 and face.get("box", [0, 0, 0, 0])[2] > 30)
    return face_count

# ------------------- API Endpoint: /predict -------------------
@app.route("/predict", methods=["POST"])
def predict():
    if "file" not in request.files:
        return jsonify({"error": "No file uploaded"}), 400
    
    file = request.files["file"]
    prediction_type = request.form.get("prediction_type", "real_vs_fake")  # default

    filename = os.path.join(UPLOAD_FOLDER, file.filename)
    file.save(filename)

    try:
        image = Image.open(filename).convert("RGB")
        image_tensor = transform(image).unsqueeze(0).to(device)
    except Exception as e:
        return jsonify({"error": "Error processing image", "details": str(e)}), 500

    # --------- CASE 1: ONLY Real/Fake Prediction ----------
    if prediction_type == "real_vs_fake":
        with torch.no_grad():
            inputs = processor(images=image, return_tensors="pt").to(device)
            outputs_realfake = realfake_detector(**inputs)
            pred_label = torch.argmax(outputs_realfake.logits, dim=1).item()

        if pred_label == 1:
            return jsonify({
                "prediction": "Real",
                "message": "Image is authentic. No further processing.",
                "image_url": url_for("static", filename=f"uploads/{file.filename}")
            })
        else:
            return jsonify({
                "prediction": "Fake",
                "message": "Image is fake, but type (Deepfake/Cheapfake) not determined in this mode.",
                "image_url": url_for("static", filename=f"uploads/{file.filename}")
            })

    # --------- CASE 2: Deepfake vs Cheapfake Analysis ----------
    elif prediction_type == "deepfake_vs_cheapfake":
        with torch.no_grad():
            deepfake_probs = torch.softmax(deepfake_model(image_tensor), dim=1)[0]
            deepfake_confidence_before = deepfake_probs[1].item() * 100  
            cheapfake_confidence_before = torch.sigmoid(cheapfake_model(image_tensor)).item() * 100  

        face_count = detect_face(filename)
        face_factor = min(face_count / 2, 1)  

        if deepfake_confidence_before <= cheapfake_confidence_before:
            adjusted_deepfake_confidence = deepfake_confidence_before * (1 + 0.3 * face_factor)
            adjusted_cheapfake_confidence = cheapfake_confidence_before * (1 - 0.3 * face_factor)
        else:
            adjusted_deepfake_confidence = deepfake_confidence_before
            adjusted_cheapfake_confidence = cheapfake_confidence_before

        fake_type = "Deepfake" if adjusted_deepfake_confidence > adjusted_cheapfake_confidence else "Cheapfake"

        return jsonify({
            "prediction": "Fake",
            "fake_type": fake_type,
            "deepfake_confidence_before": f"{deepfake_confidence_before:.2f}%",
            "deepfake_confidence_adjusted": f"{adjusted_deepfake_confidence:.2f}%",
            "cheapfake_confidence_before": f"{cheapfake_confidence_before:.2f}%",
            "cheapfake_confidence_adjusted": f"{adjusted_cheapfake_confidence:.2f}%",
            "faces_detected": face_count,
            "image_url": url_for("static", filename=f"uploads/{file.filename}")
        })

    # --------- CASE 3: Invalid prediction_type ---------
    else:
        return jsonify({"error": "Invalid prediction_type. Use 'real_vs_fake' or 'deepfake_vs_cheapfake'"}), 400

# ------------------- Heatmap Generator and API -------------------



# Flask setup

UPLOAD_FOLDER = "static/uploads"
HEATMAP_FOLDER = "static/heatmaps"
ALLOWED_EXTENSIONS = {"png", "jpg", "jpeg"}

os.makedirs(UPLOAD_FOLDER, exist_ok=True)
os.makedirs(HEATMAP_FOLDER, exist_ok=True)

def allowed_file(filename):
    return "." in filename and filename.rsplit(".", 1)[1].lower() in ALLOWED_EXTENSIONS

# Load your model
deepfake_model = load_model_from_hf("faryalnimra/DFDC-detection-model", "DFDC.pth", 2)
deepfake_model.eval()

# Choose the last Conv2D layer
target_layer = None
for name, module in deepfake_model.named_modules():
    if isinstance(module, torch.nn.Conv2d):
        target_layer = module

# Grad-CAM class
class GradCAM:
    def __init__(self, model, target_layer):
        self.model = model
        self.target_layer = target_layer
        self.gradients = None
        self.activations = None
        self._register_hooks()

    def _register_hooks(self):
        def forward_hook(module, input, output):
            self.activations = output.detach()

        def backward_hook(module, grad_in, grad_out):
            self.gradients = grad_out[0].detach()

        self.target_layer.register_forward_hook(forward_hook)
        self.target_layer.register_backward_hook(backward_hook)

    def generate(self, input_tensor, class_idx=None):
        self.model.eval()
        output = self.model(input_tensor)

        if class_idx is None:
            class_idx = torch.argmax(output, dim=1).item()

        self.model.zero_grad()
        loss = output[0, class_idx]
        loss.backward()

        gradients = self.gradients.cpu().numpy()[0]
        activations = self.activations.cpu().numpy()[0]

        weights = np.mean(gradients, axis=(1, 2))
        cam = np.zeros(activations.shape[1:], dtype=np.float32)

        for i, w in enumerate(weights):
            cam += w * activations[i, :, :]

        cam = np.maximum(cam, 0)
        cam = cv2.resize(cam, (input_tensor.size(3), input_tensor.size(2)))
        cam = cam - np.min(cam)
        cam = cam / np.max(cam)
        return cam, output

# Preprocessing
preprocess = transforms.Compose([
    transforms.Resize((224, 224)),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])

gradcam = GradCAM(deepfake_model, target_layer)

# Generate heatmap and prediction
def generate_heatmap(original_image_path, heatmap_save_path):
    img = Image.open(original_image_path).convert("RGB")
    input_tensor = preprocess(img).unsqueeze(0)

    cam, output = gradcam.generate(input_tensor)

    # Get prediction
    probabilities = torch.nn.functional.softmax(output, dim=1)[0]
    class_idx = torch.argmax(probabilities).item()
    confidence = probabilities[class_idx].item()
    label = "Fake" if class_idx == 1 else "Real"

    # Generate heatmap
    heatmap = cv2.applyColorMap(np.uint8(255 * cam), cv2.COLORMAP_JET)
    heatmap = cv2.GaussianBlur(heatmap, (7, 7), 0)
    heatmap = cv2.cvtColor(heatmap, cv2.COLOR_BGR2RGB)

    img_np = np.array(img.resize((224, 224)))

    superimposed_img = heatmap * 0.5 + img_np * 0.5
    superimposed_img = np.uint8(superimposed_img)

    Image.fromarray(superimposed_img).save(heatmap_save_path)

    return label, confidence

# Flask route
@app.route("/generate_heatmap", methods=["POST"])
def generate_heatmap_api():
    if "file" not in request.files:
        return jsonify({"error": "No file uploaded"}), 400

    file = request.files["file"]

    if file.filename == "" or not allowed_file(file.filename):
        return jsonify({"error": "Invalid file type. Allowed types are .png, .jpg, .jpeg"}), 400

    filename = secure_filename(file.filename)
    original_image_path = os.path.join(UPLOAD_FOLDER, filename)

    try:
        file.save(original_image_path)
    except Exception as e:
        return jsonify({"error": "Failed to save the file"}), 500

    heatmap_filename = f"heatmap_{filename}"
    heatmap_path = os.path.join(HEATMAP_FOLDER, heatmap_filename)

    label, confidence = generate_heatmap(original_image_path, heatmap_path)

    return jsonify({
        "original_image_url": url_for("static", filename=f"uploads/{filename}", _external=True),
        "heatmap_image_url": url_for("static", filename=f"heatmaps/{heatmap_filename}", _external=True),
        "prediction": label,
        "confidence": f"{confidence:.2f}"
    })

# To run:
# if __name__ == "__main__":
#     app.run(debug=True)







#MongoDB Atlantis from flask import Flask, request, jsonify


# MongoDB connection
client = MongoClient('mongodb+srv://fakecatcherai:[email protected]/?retryWrites=true&w=majority&appName=Cluster0')
db = client['fakecatcherDB']
users_collection = db['users']
contacts_collection = db['contacts']

def is_valid_password(password):
    if (len(password) < 8 or
        not re.search(r'[A-Z]', password) or
        not re.search(r'[a-z]', password) or
        not re.search(r'[0-9]', password) or
        not re.search(r'[!@#$%^&*(),.?":{}|<>]', password)):
        return False
    return True

@app.route('/Register', methods=['POST'])
def register():
    data = request.get_json()
    first_name = data.get('firstName')
    last_name = data.get('lastName')
    email = data.get('email')
    password = data.get('password')

    if users_collection.find_one({'email': email}):
        logging.warning(f"Attempted register with existing email: {email}")
        return jsonify({'message': 'Email already exists!'}), 400

    # βœ… Password constraints check
    if not is_valid_password(password):
        return jsonify({'message': 'Password must be at least 8 characters long and include uppercase, lowercase, number, and special character.'}), 400

    hashed_pw = generate_password_hash(password)
    users_collection.insert_one({
        'first_name': first_name,
        'last_name': last_name,
        'email': email,
        'password': hashed_pw
    })

    logging.info(f"New user registered: {first_name} {last_name}, Email: {email}")
    return jsonify({'message': 'Registration successful!'}), 201

# πŸ”΅ Login Route
@app.route('/Login', methods=['POST'])
def login():
    data = request.get_json()
    email = data.get('email')
    password = data.get('password')

    # Check if the user exists
    user = users_collection.find_one({'email': email})
    if not user or not check_password_hash(user['password'], password):
        logging.warning(f"Failed login attempt for email: {email}")
        return jsonify({'message': 'Invalid email or password!'}), 401

    logging.info(f"User logged in successfully: {email}")
    return jsonify({'message': 'Login successful!'}), 200
@app.route('/ForgotPassword', methods=['POST'])
def forgot_password():
    data = request.get_json()
    email = data.get('email')
    new_password = data.get('newPassword')
    confirm_password = data.get('confirmPassword')

    # Check if passwords match
    if new_password != confirm_password:
        logging.warning(f"Password reset failed. Passwords do not match for email: {email}")
        return jsonify({'message': 'Passwords do not match!'}), 400

    # Check if the user exists
    user = users_collection.find_one({'email': email})
    if not user:
        logging.warning(f"Password reset attempt for non-existent email: {email}")
        return jsonify({'message': 'User not found!'}), 404

    # Hash the new password and update it
    hashed_pw = generate_password_hash(new_password)
    users_collection.update_one({'email': email}, {'$set': {'password': hashed_pw}})

    logging.info(f"Password successfully reset for email: {email}")
    return jsonify({'message': 'Password updated successfully!'}), 200




   

# 🟣 Contact Form Route (React Page: Contact)
@app.route('/Contact', methods=['POST'])
def contact():
    data = request.get_json()
    email = data.get('email')
    query = data.get('query')
    message = data.get('message')

    # Check if all fields are provided
    if not email or not query or not message:
        logging.warning(f"Incomplete contact form submission from email: {email}")
        return jsonify({'message': 'All fields are required!'}), 400

    # Insert the contact data
    contact_data = {
        'email': email,
        'query': query,
        'message': message
    }
    contacts_collection.insert_one(contact_data)

    logging.info(f"Contact form submitted successfully from email: {email}")
    return jsonify({'message': 'Your message has been sent successfully.'}), 200

if __name__ == '__main__':
    app.run(debug=True)