Home

Awesome

DS-PASS

Detail-Sensitive Panoramic Annular Semantic Segmentation

PASS Dataset

GoogleDrive

New! Google Drive Download Link

BaiduYun

For Validation (Most important files):

Unfolded Panoramas for Validation, (400 images)

Annonations, (400 annotation images)

Groundtruth

There are 400 panoramas with annotations. Please use the Annotations data for evaluation.

In total, there are 1050 panoramas. Complete Panoramas:

All Unfolded Panoramas

RAW Panoramas: RAW1, RAW2, RAW3

Panorama Sequences Captured by Instrumented Vehicle

Sequence1 Sequence2 Sequence3

Panorama Sequences Captured by Mobile Robot

Sequences

Code Usage

Download the Model (model_superbest.pth) from

Trained-SwaftNet-Model BaiduYun

or

Trained-SwaftNet-Model GoogleDrive

python3.6 eval_cityscapes_color_1.py --datadir /home/kailun/Downloads/DS-PASS-master/eval_swaftnet/data/ --subset val --loadDir ../eval_swaftnet/ --loadWeights model_superbest.pth --loadModel swaftnet.py

Example segmentation

Video

Publications

If you use our code or dataset, please consider citing any of the following papers:

DS-PASS: Detail-Sensitive Panoramic Annular Semantic Segmentation through SwaftNet for Surrounding Sensing. K. Yang, X. Hu, H. Chen, K. Xiang, K. Wang, R. Stiefelhagen. In IEEE Intelligent Vehicles Symposium (IV), Las Vegas, NV, United States, October 2020. [PDF] [VIDEO]

Helping the Blind to Get through COVID-19: Social Distancing Assistant Using Real-Time Semantic Segmentation on RGB-D Video. M. Martinez, K. Yang, A. Constantinescu, R. Stiefelhagen. Sensors, 2020. [PDF]