core/predict.py

import torch
import torch.nn as nn
import torchvision.transforms as transforms
from PIL import Image
import cv2
import os


# define CNN architecture
class CustomCNNModel(nn.Module):
    def __init__(self, input_dim, num_classes):
        super(CustomCNNModel,self).__init__()
        self.input_dim = input_dim
        self.num_classes = num_classes
        
        self.conv_layers = nn.Sequential(
            nn.Conv2d(3, 32, kernel_size=3, padding =1, stride=1),
            nn.BatchNorm2d(32),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2),
            
            nn.Conv2d(32, 64, kernel_size=3, padding =1, stride=1),
            nn.BatchNorm2d(64),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2),
            
            nn.Conv2d(64, 128, kernel_size=3, padding =1, stride=1),
            nn.BatchNorm2d(128),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2),
            
            nn.Conv2d(128, 256, kernel_size=3, padding =1, stride=1),
            nn.BatchNorm2d(256),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2),         
            
        )
        self._to_linear = None
        self._get_conv_output(self.input_dim)
        
        self.fc_layers = nn.Sequential(
            nn.Linear(self._to_linear, 512),
            nn.ReLU(),
            nn.Linear(512, 128),
            nn.ReLU(),
            nn.Linear(128,self.num_classes)            
        )
        
            
    def forward(self, x):
        x = self.conv_layers(x)
        x = x.view(x.size(0),-1)
        x = self.fc_layers(x)
        return x

    def _get_conv_output(self, input_dim=128):
        with torch.no_grad():
            dummy_input = torch.zeros(1, 3,input_dim, input_dim) #batch size, no of channels, input 
            output = self.conv_layers(dummy_input)
            self._to_linear = output.view(1, -1).size(1)
            

# load CNN architecture trained weights
# index to label map
# transformation
class ImageClassifier():
    def __init__(self, model_path, class_names=None):
        self.device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
        self.model = CustomCNNModel(input_dim=128, num_classes=3).to(self.device)
        self.model.load_state_dict(torch.load(model_path, map_location = self.device))
        self.model.eval()
        if class_names is None:
            self.class_names = {0: 'person', 1: 'Dog', 2: 'Cat'}
        else:
            self.class_names = class_names
        
        self.transform = transforms.Compose([
            transforms.Resize((128, 128)),
            transforms.ToTensor(),
            transforms.Normalize(mean=[0.5, 0.5, 0.5], std = [0.5, 0.5, 0.5])
        ])

    # load image with pillow(pytorch expects pillow)
    # prediction
    # label map --> class
    # opencv, write text on our input image
    # return class, output image    
    def predict(self, image_path):
        image = Image.open(image_path).convert("RGB")
        image_tensor = self.transform(image).unsqueeze(0).to(self.device) # adds batch size to the image unsqueeze
        with torch.no_grad():
            output = self.model(image_tensor)
            ## [[0.3, 0.7, 0.9]] -> 2D tensor
            _, predicted = torch.max(output, 1)
        label = self.class_names[predicted.item()]
        
        img = cv2.imread(image_path)
        cv2.putText(img, label, (10,30), cv2.FONT_HERSHEY_SIMPLEX, 1, (0,0,255), 2)
        output_path = "output_image.jpg"
        cv2.imwrite("output_image.jpg",img)
        cwd = os.getcwd()
        output_path = os.path.join(cwd, output_path)
        
        return label, output_path
                    
            
      
        
        


        ## CNN Architecture
        ## CNN() model instance
        ## CNN.load_dict()