Skip to content
intermediateimage

PASCAL VOC Object Detection

Multi-label object classification following the PASCAL Visual Object Classes challenge (Everingham et al., IJCV 2010). Tag which of the 20 VOC object categories are present in each image, with VOC-style difficulty, truncation, and occlusion attributes.

Labels:outdoornatureurbanpeopleanimal+−

Configuration Fileconfig.yaml

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

yaml
# PASCAL VOC Object Detection Configuration
# Based on Everingham et al., IJCV 2010

annotation_task_name: "PASCAL VOC Object Detection"
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: "object_class"
    description: "Select all object classes present in the image"
    labels:
      # Vehicles
      - name: "aeroplane"
        tooltip: "Aircraft, planes"
      - name: "bicycle"
        tooltip: "Bicycles"
      - name: "boat"
        tooltip: "Boats, ships"
      - name: "bus"
        tooltip: "Buses"
      - name: "car"
        tooltip: "Cars, automobiles"
      - name: "motorbike"
        tooltip: "Motorcycles"
      - name: "train"
        tooltip: "Trains"
      # Indoor
      - name: "bottle"
        tooltip: "Bottles"
      - name: "chair"
        tooltip: "Chairs"
      - name: "diningtable"
        tooltip: "Dining tables"
      - name: "pottedplant"
        tooltip: "Potted plants"
      - name: "sofa"
        tooltip: "Sofas, couches"
      - name: "tvmonitor"
        tooltip: "TV monitors"
      # Animals
      - name: "bird"
        tooltip: "Birds"
      - name: "cat"
        tooltip: "Cats"
      - name: "cow"
        tooltip: "Cows"
      - name: "dog"
        tooltip: "Dogs"
      - name: "horse"
        tooltip: "Horses"
      - name: "sheep"
        tooltip: "Sheep"
      # Person
      - name: "person"
        tooltip: "People"

  - annotation_type: "radio"
    name: "difficulty"
    description: "Is the object difficult to detect?"
    labels:
      - name: "not_difficult"
        tooltip: "Object is clearly visible and easy to detect"
      - name: "difficult"
        tooltip: "Object is occluded, truncated, or hard to see"

  - annotation_type: "radio"
    name: "truncated"
    description: "Is the object truncated (partially outside image)?"
    labels:
      - name: "not_truncated"
        tooltip: "Object is fully visible in the image"
      - name: "truncated"
        tooltip: "Object extends beyond the image boundary"

  - annotation_type: "radio"
    name: "occluded"
    description: "Is the object occluded by another object?"
    labels:
      - name: "not_occluded"
        tooltip: "Object is not blocked by other objects"
      - name: "occluded"
        tooltip: "Object is partially hidden by other objects"

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 Datasample-data.json

json
[
  {
    "id": "voc_001",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/3/3a/Cat03.jpg",
    "context": "Select all PASCAL VOC object classes present in the image. Note if objects are truncated, occluded, or difficult to detect."
  },
  {
    "id": "voc_002",
    "image_url": "https://upload.wikimedia.org/wikipedia/commons/2/26/YellowLabradorLooking_new.jpg",
    "context": "Select all visible PASCAL VOC object classes. Indicate difficulty level for each object."
  }
]

// ... and 1 more items

Get This Design

View on GitHub

Clone or download from the repository

Quick start:

git clone https://github.com/davidjurgens/potato-showcase.git
cd potato-showcase/image/classification/pascal-voc
potato start config.yaml

Dataset & paper

Everingham et al., IJCV 2010

Official dataset ↗Read the paper ↗

Citation (BibTeX)

bibtex
@article{everingham2010pascal,
    title = "The {PASCAL} Visual Object Classes ({VOC}) Challenge",
    author = "Everingham, Mark  and Van Gool, Luc  and Williams, Christopher K. I.  and Winn, John  and Zisserman, Andrew",
    journal = "International Journal of Computer Vision",
    volume = "88",
    number = "2",
    pages = "303--338",
    year = "2010",
    publisher = "Springer",
    doi = "10.1007/s11263-009-0275-4",
    url = "https://link.springer.com/article/10.1007/s11263-009-0275-4"
}

Details

Annotation Types

multiselectradio

Domain

Computer VisionObject Detection

Use Cases

Object DetectionImage Classification

Tags

pascal-vocobject-detectionbounding-boxijcv2010benchmark

Found an issue or want to improve this design?

Open an Issue

Related Designs

MS COCO Object Detection & Categorization

Object detection annotation following the MS COCO format (Lin et al., ECCV 2014). Select all object categories present in each image and mark object locations with bounding boxes, drawn from COCO's 80 common object categories.

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

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