South Ontario Birds Family Classification Model

Overview

Dataset

Model Architecture

Training Details

• Learning Rate: 0.0001
• Batch Size: 16
• Number of Epochs: 25
• Dropout Rate: 0.6

Evaluation

• Accuracy on Test Set: The final accuracy achieved on the test set was approximately 47.04% (based on the last evaluation output).
• Other metrics like Precision, Recall, and F1-Score were also calculated and a confusion matrix was generated during evaluation.

Limitations

• The model's accuracy is moderate (around 47%), indicating that it may struggle with fine-grained classification between similar bird families.
• The dataset is limited to South Ontario, so performance on birds from other regions may vary.
• The model classifies by family, not by individual species.
• The dataset size for some bird families is relatively small, which could impact the model's ability to generalize well to those specific families.

How to Use the Model for Inference

1. Load the Model and Label Map: You will need the south_ontario_bird_model.pth state dictionary file and the label_map.json file.
   python

   1import torch 2import torch.nn as nn 3import torchvision.models as models 4import torchvision.transforms as transforms 5from PIL import Image 6import json 7 8# Define the model architecture (must match the trained model) 9def GetCleanModel(dropoutNum): 10 model_tuned = models.resnet50(weights=models.ResNet50_Weights.IMAGENET1K_V1) 11 num_ftrs_tuned = model_tuned.fc.in_features 12 model_tuned.fc = torch.nn.Sequential( 13 nn.Dropout(p=dropoutNum), 14 torch.nn.Linear(num_ftrs_tuned, 37) # 37 classes for bird families 15 ) 16 return model_tuned 17 18# Load the saved model state dictionary 19model_save_path = 'south_ontario_bird_model.pth' # Path to your saved model file 20loaded_model = GetCleanModel(dropoutNum = 0.6) # Create a new model instance with the same architecture and dropout 21loaded_model.load_state_dict(torch.load(model_save_path)) 22loaded_model.eval() # Set the model to evaluation mode 23 24# Load the label map 25label_map_save_path = 'label_map.json' # Path to your saved label map file 26with open(label_map_save_path, 'r') as f: 27 reverse_label_map = json.load(f) 28# Convert keys back to integers if they were saved as strings 29reverse_label_map = {int(k): v for k, v in reverse_label_map.items()} 30 31 32# Set device 33device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") 34loaded_model.to(device) 35 36print("Model and label map loaded successfully.")
2. Preprocess the New Image: Apply the same transformations used for the validation/test sets.
   python

   1# Define the same transformations used for validation/testing 2inference_transforms = transforms.Compose([ 3 transforms.Resize(256), 4 transforms.CenterCrop(224), 5 transforms.ToTensor(), 6 transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) 7]) 8 9# Example: Load and preprocess a new image 10new_image_path = 'path/to/your/new_bird_image.jpg' # Replace with the actual path to your new image 11 12try: 13 new_image = Image.open(new_image_path).convert('RGB') # Ensure image is RGB 14 input_tensor = inference_transforms(new_image).unsqueeze(0) # Add batch dimension 15 input_tensor = input_tensor.to(device) 16 print("Image preprocessed successfully.") 17 18except FileNotFoundError: 19 print(f"Error: New image not found at {new_image_path}") 20except Exception as e: 21 print(f"An error occurred during image preprocessing: {e}")
3. Make a Prediction: Pass the preprocessed image tensor through the loaded model.
   python

   1# Make a prediction 2if 'input_tensor' in locals(): # Check if input_tensor was created successfully 3 with torch.no_grad(): 4 output = loaded_model(input_tensor) 5 _, predicted_class_index = torch.max(output, 1); 6 7 # Convert the predicted class index back to the bird family name 8 predicted_bird_family = reverse_label_map[predicted_class_index.item()] 9 10 print(f"The predicted bird family for the new image is: {predicted_bird_family}")