Published November 17, 2026 | Version v1.0

NazmulHaqueNazim/RoadAnnoter: RoadAnnoter v1.0

Authors/Creators

Description

RoadAnnoter v1.0

A desktop tool for semantic segmentation annotation of road-scene images.

Features

  • Manual polygon drawing with undo/redo (100 steps)
  • SAM Prompt and Smart Selection using SAM / MobileSAM
  • Edge Fill: classical edge-detection + colour flood-fill (no model required)
  • Pretrained suggestions from SegFormer Cityscapes models (B0, B2, B5)
  • Own-model suggestions from TinySegNet (MobileNetV3-Small + ASPP)
  • YOLO-seg format label output with atomic file writes

Requirements

Python 3.x, PyQt5, OpenCV, NumPy Optional: torch, torchvision, transformers, segment-anything / mobile-sam

Citation

If you use this tool in your research, please cite: [add your paper citation here after publication]

Developer

Nazmul Haque

Files

NazmulHaqueNazim/RoadAnnoter-v1.0.zip

Files (67.5 kB)

Name Size
md5:60de882c6a58d3c9691a23be2d2931cd
67.5 kB Preview Download

Additional details

Related works