Skip to content
advancedimage

Cityscapes Instance Segmentation

Urban scene understanding with instance-level semantic labeling (Cordts et al., CVPR 2016). Annotate street scenes with pixel-level labels for 30 classes across vehicles, humans, construction, and nature.

Get This Design

View on GitHub(opens in a new tab)

Clone or download from the repository

Quick start:

bash
git clone https://github.com/davidjurgens/potato-showcase.git
cd potato-showcase/image/segmentation/cityscapes
potato start config.yaml

Configuration File

This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.

yaml
# Cityscapes Instance Segmentation Configuration
# Based on Cordts et al., CVPR 2016

annotation_task_name: "Cityscapes Urban Scene Segmentation"
task_dir: "."

data_files:
  - "sample-data.json"

item_properties:
  id_key: "id"
  text_key: "image_url"
  context_key: "context"

user_config:
  allow_all_users: true

annotation_schemes:
  - annotation_type: "multiselect"
    name: "flat_surfaces"
    description: "Select visible flat surface classes"
    labels:
      - name: "road"
        tooltip: "Drivable road surface"
      - name: "sidewalk"
        tooltip: "Pedestrian sidewalk"
      - name: "parking"
        tooltip: "Parking area"
      - name: "rail_track"
        tooltip: "Rail tracks"

  - annotation_type: "multiselect"
    name: "humans"
    description: "Select visible human classes"
    labels:
      - name: "person"
        tooltip: "Pedestrian or person"
      - name: "rider"
        tooltip: "Person on bicycle/motorcycle"

  - annotation_type: "multiselect"
    name: "vehicles"
    description: "Select visible vehicle classes"
    labels:
      - name: "car"
        tooltip: "Cars, SUVs"
      - name: "truck"
        tooltip: "Trucks"
      - name: "bus"
        tooltip: "Buses"
      - name: "train"
        tooltip: "Trains, trams"
      - name: "motorcycle"
        tooltip: "Motorcycles, mopeds"
      - name: "bicycle"
        tooltip: "Bicycles"

  - annotation_type: "multiselect"
    name: "construction"
    description: "Select visible construction elements"
    labels:
      - name: "building"
        tooltip: "Buildings"
      - name: "wall"
        tooltip: "Walls"
      - name: "fence"
        tooltip: "Fences"
      - name: "bridge"
        tooltip: "Bridges"
      - name: "tunnel"
        tooltip: "Tunnels"

  - annotation_type: "multiselect"
    name: "nature"
    description: "Select visible nature elements"
    labels:
      - name: "vegetation"
        tooltip: "Trees, bushes, plants"
      - name: "terrain"
        tooltip: "Grass, soil"
      - name: "sky"
        tooltip: "Sky"

  - annotation_type: "multiselect"
    name: "objects"
    description: "Select visible objects"
    labels:
      - name: "pole"
        tooltip: "Poles, posts"
      - name: "traffic_sign"
        tooltip: "Traffic signs"
      - name: "traffic_light"
        tooltip: "Traffic lights"

interface_config:
  item_display_format: "<img src='{{text}}' style='max-width:100%; max-height:500px;'/><br/><small>{{context}}</small>"

output_annotation_format: "json"
output_annotation_dir: "annotations"

Sample Data

json
[
  {
    "id": "city_001",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/9/9e/Frankfurt_Skyline.jpg/1200px-Frankfurt_Skyline.jpg",
    "context": "Urban street scene. Label all visible objects according to Cityscapes categories: vehicles, humans, construction, nature, and objects."
  },
  {
    "id": "city_002",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/5/5a/Spyker_C8_Spyder.jpg/1200px-Spyker_C8_Spyder.jpg",
    "context": "Street scene with vehicles. Annotate all semantic categories visible."
  }
]

// ... and 1 more items

Dataset & paper

Cordts et al., CVPR 2016

Official dataset(opens in a new tab)Read the paper(opens in a new tab)

Citation (BibTeX)

bibtex
@inproceedings{cordts2016cityscapes,
    title = "The {Cityscapes} Dataset for Semantic Urban Scene Understanding",
    author = "Cordts, Marius  and Omran, Mohamed  and Ramos, Sebastian  and Rehfeld, Timo  and Enzweiler, Markus  and Benenson, Rodrigo  and Franke, Uwe  and Roth, Stefan  and Schiele, Bernt",
    booktitle = "Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition",
    pages = "3213--3223",
    year = "2016",
    url = "https://arxiv.org/abs/1604.01685"
}

Details

Annotation Types

multiselect

Domain

Computer VisionAutonomous Driving

Use Cases

Semantic SegmentationInstance SegmentationUrban Scene Understanding

Tags

cityscapessegmentationautonomous-drivingurbancvpr2016

Found an issue or want to improve this design?

Open an Issue(opens in a new tab)

Related Designs

BDD100K: Diverse Driving Dataset for Autonomous Driving

BDD100K is a 100,000-video driving dataset from UC Berkeley (Yu et al., CVPR 2020) spanning detection, segmentation, lane marking, and tracking. This Potato config reproduces its driving-scene labeling tasks.

multiselectradio

ADE20K: Scene Parsing and Semantic Segmentation Dataset

ADE20K is a scene parsing dataset from MIT CSAIL (Zhou et al., CVPR 2017) with dense pixel labels for objects, stuff, and parts. This Potato config adapts it to an image-level category-tagging task over a curated subset of common ADE20K categories.

multiselectradio

CelebA: Large-Scale Face Attributes Dataset

CelebA is a face dataset of 202,599 celebrity images, each labeled with 40 binary attributes, identity, and 5 landmark points (Liu et al., ICCV 2015). This Potato config reproduces the 40-attribute multiselect task.

multiselect