File size: 3,438 Bytes
8c30b41
f034887
 
f658766
f23fa83
923a637
f658766
505b320
f658766
f034887
 
 
 
505b320
 
 
 
 
 
 
f658766
 
505b320
f658766
f034887
 
 
 
 
505b320
 
f034887
 
f23fa83
 
 
 
 
 
 
 
 
 
 
f658766
f23fa83
31daa87
f23fa83
 
31daa87
 
 
f034887
505b320
f034887
 
 
 
 
31daa87
505b320
31daa87
505b320
 
 
 
 
 
 
31daa87
f034887
31daa87
f034887
505b320
f23fa83
f034887
f23fa83
31daa87
f034887
31daa87
f034887
31daa87
f034887
31daa87
 
f034887
31daa87
f034887
f658766
f034887
 
f658766
923a637
f658766
923a637
31daa87
 
 
f23fa83
f034887
f23fa83
923a637
f658766
923a637
f658766
 
 
c03cdfb
 
8c30b41
923a637
31daa87
f034887
8c30b41
 
923a637
 
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
import gradio as gr
from transformers import pipeline
from PIL import Image, ExifTags
import numpy as np
import cv2

# ----------------------------
# MODEL
# ----------------------------
try:
    hf_detector = pipeline("image-classification", model="umm-maybe/AI-image-detector")
except Exception as e:
    hf_detector = None
    print("HF AI-detector gagal dimuat:", e)

try:
    general_model = pipeline("image-classification", model="google/vit-base-patch16-224")
except Exception as e:
    general_model = None
    print("General classifier gagal dimuat:", e)

# ----------------------------
# ANALISIS LOKAL
# ----------------------------
def calculate_blur(image):
    gray = np.array(image.convert("L"))
    return cv2.Laplacian(gray, cv2.CV_64F).var()

def calculate_noise(image):
    gray = np.array(image.convert("L"), dtype=np.float32)
    noise_std = np.std(gray - np.mean(gray))
    return noise_std

def has_camera_exif(image):
    try:
        exif = image._getexif()
        if exif:
            for tag, value in exif.items():
                decoded = ExifTags.TAGS.get(tag, tag)
                if decoded in ["Make", "Model"]:
                    return True
    except:
        return False
    return False

# ----------------------------
# DETEKSI HYBRID WEIGHTED
# ----------------------------
def detect_image(image):
    hf_score = 0
    general_score = 0
    local_score = 0

    # -------- HF AI-detector --------
    if hf_detector:
        try:
            result = hf_detector(image)
            label = result[0]['label'].lower()
            conf = result[0]['score'] * 100
            hf_score = conf if any(x in label for x in ["fake", "ai", "artificial"]) else 0
        except:
            hf_score = 0

    # -------- General model --------
    if general_model:
        try:
            result2 = general_model(image)
            label2 = result2[0]['label'].lower()
            conf2 = result2[0]['score'] * 100
            general_score = conf2 if any(x in label2 for x in ["anime","cartoon","illustration","maya"]) else 0
        except:
            general_score = 0

    # -------- Analisis lokal --------
    blur_score = calculate_blur(image)
    noise_score = calculate_noise(image)
    exif_present = has_camera_exif(image)
    local_score = 0
    if blur_score < 100 or noise_score < 10:
        local_score += 50
    if not exif_present:
        local_score += 10

    # -------- Weighted Score --------
    final_score = hf_score*0.6 + general_score*0.25 + local_score*0.15

    if final_score > 50:
        final_result = "🤖 AI Detected"
    else:
        final_result = "✅ Foto Asli"

    output = f"""
### Hasil Deteksi:
{final_result}

**Weighted Skor:** {final_score:.2f}
**HF AI-detector:** {result[0]['label']} ({result[0]['score']*100:.2f}%)
**General Model:** {result2[0]['label']} ({result2[0]['score']*100:.2f}%)
**Blur Score:** {blur_score:.2f}
**Noise Score:** {noise_score:.2f}
**Metadata Kamera:** {'Ada' if exif_present else 'Tidak Ada'}
"""
    return output

# ----------------------------
# Gradio Interface
# ----------------------------
iface = gr.Interface(
    fn=detect_image,
    inputs=gr.Image(type="pil"),
    outputs="markdown",
    title="AI vs Foto Asli Detector (Weighted Hybrid)",
    description="Unggah gambar, sistem hybrid akan mendeteksi apakah gambar kemungkinan besar asli atau dihasilkan AI."
)

if __name__ == "__main__":
    iface.launch()