MCS Dataset

Overview

MCS Dataset focuses on semantic understanding of vehicles and curbside in an urban environment.

Features

  • Annotations
    Dense semantic segmentation
    Pixel-level semantic labeling

  • Volume
    20 000 annotated images with fine annotations

  • Complexity
    20 classes

  • Diversity
    Daytime
    Good/medium weather conditions
    Manually selected frames
    95% synthetic
    5% real world data

Examples

Vehicle Annotation Example

[{
“camera_tag”: “MainCamera”,
“type”: “Hatchback”,
“instance_id”: 1,
“truncated”: 0.0,
“occluded”: 2,
“rotationY”: -0.008393635042011738,
“thetaRayY”: -0.4797818958759308,
“alphaY”: 0.4713882505893707,
“thetaRayX”: -0.15779972076416017,
“rotationX”: -0.09106016159057617,
“alphaX”: 0.06673955917358399,
“thetaRayZ”: 2.7941994667053224,
“rotationZ”: 1.5703239440917969,
“alphaZ”: -1.2238755226135255
},{
“camera_tag”: “MainCamera”,
“type”: “Pickup truck”,
“instance_id”: 2,
“truncated”: 0.0,
“occluded”: 0,
“rotationY”: 1.5707961320877076,
“thetaRayY”: -0.10266316682100296,
“alphaY”: 1.6734592914581299,
“thetaRayX”: -0.0839700698852539,
“rotationX”: 0.23560333251953126,
“alphaX”: 0.31957340240478518,
“thetaRayZ”: 2.182987689971924,
“rotationZ”: 0.0,
“alphaZ”: -2.182987689971924
},{

}]

Availability

September 2021

License

MCS Dataset is made freely available to academic and non-academic entities for non-commercial purposes such as academic research, teaching, scientific publications, or personal experimentation.