Project Overview

This project addresses the challenge of training a deep learning model for image classification on a dataset with eight distinct classes. The primary issue is the domain shift between the training and test datasets, where training images have varied backgrounds, but test images have a uniform background. This discrepancy can ostacolate the model's performance in generalizing from training to test data.

Proposed Solution

To tackle this challenge, we propose a solution that involves:

• Background Removal: Using an additional neural network to predict bounding boxes for the images.
• Bounding Box Prediction: Training a model to detect bounding boxes for test set images, as bounding boxes are only available for the training set.
• Focus on Objects: By concentrating on the objects within the bounding boxes and ignoring the backgrounds, the model can better generalize and accurately classify images regardless of the background.

Key Idea

The key idea in our approach is the use of background removal to align the training and test data distributions. This strategy reduces the impact of domain shift and enhances the model's ability to accurately classify images.

Steps and Methodology

• Data Preparation:
  • Load and preprocess the dataset.
  • Use bounding box information from the training set to focus on objects and remove backgrounds.
• Model Training:
  • Train a neural network to predict bounding boxes for test set images.
  • Use these predicted bounding boxes to remove backgrounds from test set images.
• Image Classification:

• Train the main image classification model on the background-removed images.
• Evaluate the model's performance on the test set.

Requirements

Python 3.x Libraries: TensorFlow or PyTorch, OpenCV, scikit-learn, numpy, matplotlib