File size: 10,276 Bytes
12ad63a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1f3a1d8
 
 
 
 
 
 
 
 
 
12ad63a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
11522ca
 
12ad63a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""
FastAPI application for plant image classification using ViT-ConvNext hybrid model.
Provides two endpoints:
1. /predict/upload - Upload image file directly
2. /predict/url - Provide image URL
Both return top 5 predictions with confidence scores.
"""

import torch
import torch.nn as nn
import timm
from fastapi import FastAPI, UploadFile, File, HTTPException
from fastapi.responses import JSONResponse
from pydantic import BaseModel, HttpUrl
from PIL import Image
from torchvision import transforms
import io
import requests
from typing import List, Dict
import ast

# ============== Model Architecture ==============
class CBAMBlock(nn.Module):
    """CBAM (Channel + Spatial Attention) Block"""
    def __init__(self, channels, reduction=16, spatial_kernel=7):
        super(CBAMBlock, self).__init__()
        # Channel attention
        self.channel_att = nn.Sequential(
            nn.AdaptiveAvgPool2d(1),
            nn.Conv2d(channels, channels // reduction, 1, bias=False),
            nn.ReLU(inplace=True),
            nn.Conv2d(channels // reduction, channels, 1, bias=False),
            nn.Sigmoid()
        )
        # Spatial attention
        self.spatial_att = nn.Sequential(
            nn.Conv2d(2, 1, kernel_size=spatial_kernel, padding=spatial_kernel // 2, bias=False),
            nn.Sigmoid()
        )

    def forward(self, x):
        # Channel attention
        ca = self.channel_att(x)
        x = x * ca
        # Spatial attention
        sa = torch.cat([
            torch.mean(x, dim=1, keepdim=True), 
            torch.max(x, dim=1, keepdim=True)[0]
        ], dim=1)
        sa = self.spatial_att(sa)
        x = x * sa
        return x


class ViTCNNHybrid(nn.Module):
    """Hybrid model combining Swin Transformer and ConvNeXt with gated fusion"""
    def __init__(self, num_classes, use_cbam=True):
        super(ViTCNNHybrid, self).__init__()
        
        # Swin Transformer branch
        self.vit = timm.create_model(
            'swin_tiny_patch4_window7_224', pretrained=False, num_classes=0, drop_rate=0.3
        )
        self.vit_out_features = 768
        
        # ConvNeXt-Tiny branch
        self.cnn = timm.create_model(
            'convnext_tiny', pretrained=False, num_classes=0, drop_rate=0.3, global_pool=''
        )
        self.cnn_out_features = 768
        self.cnn_pool = nn.AdaptiveAvgPool2d((7, 7))
        
        # Gates for gated fusion
        self.vit_gate = nn.Sequential(
            nn.AdaptiveAvgPool2d(1),
            nn.Conv2d(self.vit_out_features, self.vit_out_features // 16, 1),
            nn.ReLU(inplace=True),
            nn.Conv2d(self.vit_out_features // 16, self.vit_out_features, 1),
            nn.Sigmoid()
        )
        self.cnn_gate = nn.Sequential(
            nn.AdaptiveAvgPool2d(1),
            nn.Conv2d(self.cnn_out_features, self.cnn_out_features // 16, 1),
            nn.ReLU(inplace=True),
            nn.Conv2d(self.cnn_out_features // 16, self.cnn_out_features, 1),
            nn.Sigmoid()
        )
        
        self.match_dim = nn.Conv2d(self.vit_out_features, self.cnn_out_features, 1)

        # Learnable α for dynamic fusion
        self.alpha_param = nn.Parameter(torch.tensor(0.5))

        # Fusion layers
        fusion_layers = [
            nn.Conv2d(self.cnn_out_features, 256, kernel_size=1),
            nn.BatchNorm2d(256),
            nn.ReLU(inplace=True),
            nn.Dropout(0.3)
        ]
        if use_cbam:
            fusion_layers.append(CBAMBlock(256))
        fusion_layers.append(nn.AdaptiveAvgPool2d((1, 1)))
        self.fusion = nn.Sequential(*fusion_layers)
        
        # Classification head
        self.fc = nn.Sequential(
            nn.Linear(256, 512),
            nn.ReLU(inplace=True),
            nn.Dropout(0.4),
            nn.Linear(512, num_classes)
        )
    
    def forward(self, x):
        # ViT branch
        vit_out = self.vit(x)
        vit_out = vit_out.view(-1, self.vit_out_features, 1, 1).expand(-1, -1, 7, 7)
        vit_out = vit_out * self.vit_gate(vit_out)
        
        # CNN branch
        cnn_out = self.cnn(x)
        cnn_out = self.cnn_pool(cnn_out)
        cnn_out = cnn_out * self.cnn_gate(cnn_out)

        # Dynamic Fusion
        alpha = torch.sigmoid(self.alpha_param)
        combined = alpha * vit_out + (1 - alpha) * cnn_out
        
        combined = self.fusion(combined)
        combined = combined.view(combined.size(0), -1)
        out = self.fc(combined)
        return out


# ============== FastAPI Setup ==============
app = FastAPI(
    title="Plant Classification API",
    description="API for classifying plant images using ViT-ConvNext hybrid model",
    version="1.0.0"
)

# Global variables
model = None
class_names = None
device = None
transform = None


class ImageURL(BaseModel):
    """Request model for URL-based prediction"""
    url: HttpUrl


class Prediction(BaseModel):
    """Single prediction result"""
    class_name: str
    confidence: float


class PredictionResponse(BaseModel):
    """Response model containing top 5 predictions"""
    predictions: List[Prediction]


def load_class_names(file_path: str = "class.txt") -> List[str]:
    """Load class names from file"""
    with open(file_path, 'r', encoding='utf-8') as f:
        content = f.read()
        # Parse the list from the file
        classes = ast.literal_eval(content.split('Classes: ')[1])
        return classes


def get_transform():
    """Get image preprocessing transform matching training pipeline"""
    return transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
    ])


@app.on_event("startup")
async def startup_event():
    """Load model and class names on startup"""
    global model, class_names, device, transform
    
    print("Loading class names...")
    class_names = load_class_names()
    num_classes = len(class_names)
    print(f"Loaded {num_classes} classes")
    
    print("Loading model...")
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    print(f"Using device: {device}")
    
    # Initialize model
    model = ViTCNNHybrid(num_classes=num_classes, use_cbam=True)
    
    # Load trained weights
    checkpoint = torch.load("pbl6_model.pth", map_location=device)
    
    # Handle DataParallel checkpoint (remove 'module.' prefix)
    if list(checkpoint.keys())[0].startswith('module.'):
        from collections import OrderedDict
        new_checkpoint = OrderedDict()
        for k, v in checkpoint.items():
            name = k[7:]  # remove 'module.' prefix
            new_checkpoint[name] = v
        checkpoint = new_checkpoint
    
    model.load_state_dict(checkpoint)
    model.to(device)
    model.eval()
    
    # Initialize transform
    transform = get_transform()
    
    print("Model loaded successfully!")


def predict_image(image: Image.Image) -> List[Dict[str, float]]:
    """
    Perform prediction on a PIL Image
    
    Args:
        image: PIL Image object
        
    Returns:
        List of top 5 predictions with class names and confidence scores
    """
    # Convert to RGB if needed
    if image.mode != 'RGB':
        image = image.convert('RGB')
    
    # Preprocess image
    img_tensor = transform(image).unsqueeze(0).to(device)
    
    # Inference
    with torch.no_grad():
        outputs = model(img_tensor)
        probabilities = torch.nn.functional.softmax(outputs, dim=1)
    
    # Get top 5 predictions
    top5_prob, top5_idx = torch.topk(probabilities, 5, dim=1)
    top5_prob = top5_prob.cpu().numpy()[0]
    top5_idx = top5_idx.cpu().numpy()[0]
    
    # Format results
    predictions = []
    for idx, prob in zip(top5_idx, top5_prob):
        predictions.append({
            "class_name": class_names[idx],
            "confidence": float(prob)
        })
    
    return predictions


@app.post("/predict/upload", response_model=PredictionResponse)
async def predict_upload(file: UploadFile = File(...)):
    """
    Classify a plant image uploaded as a file
    
    Args:
        file: Image file (JPEG, PNG, etc.)
        
    Returns:
        Top 5 predictions with class names and confidence scores
    """
    try:
        # Read image file
        contents = await file.read()
        image = Image.open(io.BytesIO(contents))
        
        # Get predictions
        predictions = predict_image(image)
        
        return JSONResponse(content={"predictions": predictions})
        
    except Exception as e:
        raise HTTPException(status_code=400, detail=f"Error processing image: {str(e)}")


@app.post("/predict/url", response_model=PredictionResponse)
async def predict_url(image_url: ImageURL):
    """
    Classify a plant image from a URL
    
    Args:
        image_url: JSON body containing the image URL
        
    Returns:
        Top 5 predictions with class names and confidence scores
    """
    try:
        # Download image from URL with longer timeout for large images
        response = requests.get(str(image_url.url), timeout=30)
        response.raise_for_status()
        
        # Open image
        image = Image.open(io.BytesIO(response.content))
        
        # Get predictions
        predictions = predict_image(image)
        
        return JSONResponse(content={"predictions": predictions})
        
    except requests.exceptions.RequestException as e:
        raise HTTPException(status_code=400, detail=f"Error downloading image: {str(e)}")
    except Exception as e:
        raise HTTPException(status_code=400, detail=f"Error processing image: {str(e)}")


@app.get("/")
async def root():
    """Health check endpoint"""
    return {
        "message": "Plant Classification API",
        "status": "running",
        "model_loaded": model is not None,
        "num_classes": len(class_names) if class_names else 0
    }


@app.get("/health")
async def health():
    """Detailed health check"""
    return {
        "status": "healthy",
        "model": "loaded" if model is not None else "not loaded",
        "device": str(device) if device else "unknown",
        "classes": len(class_names) if class_names else 0
    }