Image Classification

Multi-label image classification with quality assessment for computer vision datasets.

Configuration Fileconfig.yaml

annotation_task_name: "Image Classification"

port: 8000

# Image display settings
image:
  enabled: true
  max_width: 800
  max_height: 600
  zoom_enabled: true
  preserve_aspect_ratio: true

# Data configuration
data_files:
  - "data/images.json"

# Annotation schemes
annotation_schemes:
  # Primary category (single choice)
  - annotation_type: radio
    name: primary_category
    description: "What is the main subject of this image?"
    labels:
      - name: Person/People
        key_value: "1"
      - name: Animal
        key_value: "2"
      - name: Vehicle
        key_value: "3"
      - name: Building/Architecture
        key_value: "4"
      - name: Nature/Landscape
        key_value: "5"
      - name: Food
        key_value: "6"
      - name: Object/Product
        key_value: "7"
      - name: Text/Document
        key_value: "8"
      - name: Other
        key_value: "9"
    sequential_key_binding: true

  # Additional tags (multi-label)
  - annotation_type: multiselect
    name: tags
    description: "Select all applicable tags"
    labels:
      - Indoor
      - Outdoor
      - Daytime
      - Nighttime
      - Close-up
      - Wide shot
      - Multiple subjects
      - Text visible
      - Watermark present

  # Scene type
  - annotation_type: radio
    name: scene_type
    description: "What type of scene is this?"
    labels:
      - Natural
      - Urban
      - Studio/Staged
      - Screenshot
      - Artwork/Illustration
      - Mixed/Unclear

  # Image quality assessment
  - annotation_type: likert
    name: image_quality
    description: "Rate the overall image quality"
    size: 5
    min_label: "Very poor"
    max_label: "Excellent"

  # Usability for training
  - annotation_type: radio
    name: usable_for_training
    description: "Is this image suitable for model training?"
    labels:
      - Yes - High quality
      - Yes - Acceptable
      - No - Quality issues
      - No - Content issues
      - Unsure

  # Quality issues (if any)
  - annotation_type: multiselect
    name: quality_issues
    description: "Select any quality issues (if applicable)"
    labels:
      - Blurry
      - Too dark
      - Too bright
      - Low resolution
      - Cropped badly
      - Watermark obstructs content
      - NSFW content
      - Duplicate
      - No issues

# User settings
allow_all_users: true
instances_per_annotator: 200
annotation_per_instance: 2

# Output
output_annotation_dir: "annotation_output/"
output_annotation_format: "json"

Get This Design

This design is available in our showcase. Copy the configuration below to get started.

Quick start:

# Create your project folder
mkdir image-classification
cd image-classification
# Copy config.yaml from above
potato start config.yaml

Details

Annotation Types

multiselectradiolikert

Domain

Computer Vision

Use Cases

classificationimage labelingdataset creation

Related Designs

ADE20K Semantic Segmentation

Comprehensive scene parsing with 150 semantic categories (Zhou et al., CVPR 2017). Annotate indoor and outdoor scenes with pixel-level labels covering objects, parts, and stuff classes.

multiselectradio

AnnoMI Counselling Dialogue Annotation

Annotation of motivational interviewing counselling dialogues based on the AnnoMI dataset. Annotators label therapist and client utterances for MI techniques (open questions, reflections, affirmations) and client change talk (sustain talk, change talk), with quality ratings for therapeutic interactions.

radiomultiselect

Audio Transcription Review

Review and correct automatic speech recognition transcriptions with waveform visualization.

likertmultiselect