ICPR-24 Competition on Safe Segmentation of Drive Scenes in Unstructured Traffic and Adverse Weather Condition

This repository contains the datasets related to safe segmentation challenge for ICPR-24 Competition on Safe Segmentation of Drive Scenes in Unstructured Traffic and Adverse Weather Condition, ICPR 2024 . For more details, please visit https://mobility.iiit.ac.in/safe-seg-24/

Installing requirements

Firstly install the required libraries using

pip install -r requirements.txt

Target datasets:

For using first add helpers/ to $PYTHONPATH

export PYTHONPATH="${PYTHONPATH}:helpers/"

Dataset Structure

The structure is similar to the IDD dataset. That is:

gtFine/{split}/{drive_no}/{img_id}_gtFine_mask.json for ground truths
leftImg8bit/{split}/{drive_no}/{img_id}_rgb.png for image frames

Semantic Segmentation

Furthermore for training, label masks needs to be generated as described below resulting in the following files:

gtFine/{split}/{drive_no}/{img_id}_gtFine_labellevel3Ids.png

Labels

See helpers/anue_labels.py

Generate Label Masks (for training/evaluation) (Semantic/Instance/Panoptic Segmentation)

python preperation/createLabels_iddaw.py --datadir $IDDAW --id-type $IDTYPE --color [True|False] --num-workers $C

• IDDAW is the path to the IDD-AW dataset
• IDTYPE can be id, csId, csTrainId, level3Id, level2Id, level1Id.
• color True generates the color masks
• C is the number of threads to run in parallel

The generated files:

• _gtFine_labelLevel3Ids.png will be used for semantic segmentation

Viewer

First generate label masks as described above. To view the ground truths / prediction masks at different levels of heirarchy use:

python viewer/viewer.py ---datadir $ANUE

• ANUE has the folder path to the dataset or prediction masks with similar file/folder structure as dataset.

TODO: Make the color map more sensible.