This task has been prepared for the Computer Vision Engineer Virtual Internship Program in Intern2Grow.
Your task is to build a program that can classify product images into one of four categories: T-shirt, Trousers, Shoes, Glasses.
-
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).
-
Model Selection and Training: Choose an appropriate machine learning model for this classification task. Justify your choice. Train your model using the training data.
-
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.
-
Model Optimization: If necessary, perform hyperparameter tuning to optimize your model's performance.
-
Prediction: Finally, use your trained model to classify new product images.
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.