Dataset parse

[pranavreddy23]

Pull Request Summary

This pull request introduces a set of scripts to standardize the parsing and visualization of the nuImages dataset. A dedicated directory, AutoSpeed/NuImages/, is created to house all tools necessary for generating training data.

---

Key Features

Annotation Generation (create_bbox_json.py)

• Parses the raw nuImages dataset.
• Generates structured JSON annotation files for each image.
• Organizes outputs by data split (train / val / test) and camera.
• Automates the creation of bounding box labels required for model training.

---

Visualization (test_Visualization.py)

• A utility script to visualize the generated bounding box annotations.
• Overlays annotations on corresponding images.
• Helps verify the correctness of annotations before training.

---

Documentation (README.md)

• A detailed guide explaining the required directory structure for the nuImages dataset.
• Step-by-step instructions on how to use:
  • The annotation generation script.
  • The visualization tool.