Update convert_segment_masks_to_yolo_seg to support custom datasets

[RizwanMunawar]

This PR aims to resolve issue #15098.

@glenn-jocher

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Documentation and code improvements for the convert_segment_masks_to_yolo_seg function.

📊 Key Changes

• Clarified documentation for the convert_segment_masks_to_yolo_seg function in both the code and usage guidelines.
• Updated 'classes' parameter description to indicate it represents the total number of classes in any dataset, not just COCO.
• Modified pixel-to-class mapping in the code to dynamically use the classes parameter value.

🎯 Purpose & Impact

• Improved Clarity: The revised documentation helps users understand that the classes parameter should reflect the total classes in their specific dataset, not just COCO. 📝
• Code Flexibility: By making the pixel-to-class mapping dynamic based on the classes parameter, the function is now more adaptable to different datasets, improving usability and reducing potential errors. 🚀

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Minor updates to documentation and code to clarify that the 'classes' parameter represents the total number of classes in the dataset.

📊 Key Changes

• Documentation: Updated comments in both the simple-utilities.md file and the converter.py file to make it clear that the 'classes' parameter is the total number of classes in the dataset.
• Code: Adjusted the pixel-to-class mapping logic to use the 'classes' parameter instead of a hardcoded value.

🎯 Purpose & Impact

• Clarity: Provides better understanding for users by clarifying the purpose of the 'classes' parameter, ensuring that users know they should set this parameter to the total number of classes in their specific dataset.
• Flexibility: Makes the function more adaptable to datasets with differing numbers of classes, enhancing its usability across various datasets beyond just COCO.

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 1 line in your changes missing coverage. Please review.

Project coverage is 66.57%. Comparing base (3e4a581) to head (9ab735d).

Files	Patch %	Lines
ultralytics/data/converter.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #15176      +/-   ##
==========================================
+ Coverage   66.53%   66.57%   +0.03%     
==========================================
  Files         129      129              
  Lines       16944    16944              
==========================================
+ Hits        11274    11280       +6     
+ Misses       5670     5664       -6     

Flag	Coverage Δ
Benchmarks	34.30% <0.00%> (+0.03%)	⬆️
GPU	35.34% <0.00%> (ø)
Tests	63.01% <0.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[PythoneerSamurai]

There are issues in this code. If I only have one class in my entire dataset, represented by pixel value 255, the dictionary will not map 255 to class-id 0, rather it would map 1 to class-id 0 (pixel_to_class_mapping = {i + 1: i for i in range(classes)}), resultantly the same issue will arise as mentioned in my first post. That's the thing my code handles, it automatically detects all classes, and then notifies the class-id and pixel-values to the users so that they can understand which pixel value represents which class.

[glenn-jocher]

PR merged!