Computer Vision Task: Product Picture Classification

This task has been prepared for the Computer Vision Engineer Virtual Internship Program in Intern2Grow.

Objective

Your task is to build a program that can classify product images into one of four categories: T-shirt, Trousers, Shoes, Glasses.

Task Breakdown

  1. Data Preparation: Import the necessary libraries. Acquire a suitable dataset from Roboflow Universe, a platform offering a large collection of open-source computer vision datasets. The dataset should contain images of products and their corresponding labels. Preprocess the images if necessary (e.g., resize, normalize).

  2. Model Selection and Training: Choose an appropriate machine learning model for this classification task. Justify your choice. Train your model using the training data.

  3. Model Evaluation: Use a validation set to evaluate the performance of your model. You may use appropriate metrics for classification tasks such as Precision, Recall, F1-Score, and Accuracy.

  4. Model Optimization: If necessary, perform hyperparameter tuning to optimize your model's performance.

  5. Prediction: Finally, use your trained model to classify new product images.

Deliverable

Submit a report detailing your approach, methodology, and results. Your report should include:

  • An explanation of your data preparation steps.
  • The machine learning model you chose and why you chose it.
  • The results of your model evaluation, including the metrics you used.
  • A discussion of any challenges you faced and how you overcame them.

Also, submit your code files along with your report. Your code should be well-commented and easy to understand.