CUB-200-2011: Fine-Grained Bird Species Dataset
CUB-200-2011 is a fine-grained bird image dataset of 11,788 photos across 200 species, each labeled with parts, attributes, and a bounding box (Wah et al., 2011). This Potato config reproduces the labeling task.
About this dataset
CUB-200-2011 (Caltech-UCSD Birds-200-2011) is a fine-grained image classification dataset built by Catherine Wah, Steve Branson, Peter Welinder, Pietro Perona, and Serge Belongie at Caltech and UCSD. It was released in 2011 as Caltech technical report CNS-TR-2011-001 and remains a standard benchmark for fine-grained visual categorization.
The dataset holds 11,788 bird photographs spread over 200 species, split into 5,994 training images and 5,794 test images. Each image carries detailed annotations collected and filtered through Amazon Mechanical Turk.
Every image is labeled with one species category, 15 part locations (such as bill, crown, and tail), 312 binary visual attributes (for example bill shape, wing pattern, and breast color), and one bounding box around the bird. The fine-grained nature of the task comes from telling apart visually similar species that differ only in small details.
The Potato config below reproduces this labeling task: a radio scheme for the bird family, a multiselect scheme for visible visual attributes drawn from the CUB attribute set, a radio scheme for visibility/occlusion, and a free-text field for notes.
- Images
- 11,788
- Species categories
- 200
- Train / test split
- 5,994 / 5,794
- Part locations per image
- 15
- Binary attributes per image
- 312
- Bounding boxes per image
- 1
Configuration Fileconfig.yaml
This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.
# CUB-200-2011 Fine-Grained Bird Classification Configuration
# Based on Wah et al., 2011
annotation_task_name: "CUB-200 Bird Species Classification"
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: "radio"
name: "bird_family"
description: "Select the bird family"
labels:
- name: "warbler"
tooltip: "Small songbirds including warblers"
- name: "sparrow"
tooltip: "Sparrows and related species"
- name: "vireo"
tooltip: "Vireos"
- name: "woodpecker"
tooltip: "Woodpeckers and related"
- name: "flycatcher"
tooltip: "Flycatchers"
- name: "wren"
tooltip: "Wrens"
- name: "gull"
tooltip: "Gulls and terns"
- name: "tern"
tooltip: "Terns"
- name: "blackbird"
tooltip: "Blackbirds and orioles"
- name: "other"
tooltip: "Other bird families"
- annotation_type: "multiselect"
name: "visual_attributes"
description: "Select visible attributes"
labels:
- name: "has_bill_shape_hooked"
tooltip: "Bill is hooked"
- name: "has_bill_shape_cone"
tooltip: "Bill is cone-shaped"
- name: "has_wing_pattern_striped"
tooltip: "Wings have striped pattern"
- name: "has_wing_pattern_solid"
tooltip: "Wings are solid colored"
- name: "has_breast_pattern_spotted"
tooltip: "Breast has spots"
- name: "has_breast_color_yellow"
tooltip: "Breast is yellow"
- name: "has_breast_color_white"
tooltip: "Breast is white"
- name: "has_crown_color_black"
tooltip: "Crown is black"
- name: "has_eye_color_red"
tooltip: "Eyes are red"
- name: "has_tail_shape_forked"
tooltip: "Tail is forked"
- annotation_type: "radio"
name: "visibility"
description: "How visible is the bird?"
labels:
- name: "fully_visible"
tooltip: "Bird is completely visible"
- name: "partially_occluded"
tooltip: "Bird is partially hidden"
- name: "mostly_occluded"
tooltip: "Most of bird is hidden"
- annotation_type: "text"
name: "species_name"
description: "Enter the specific species name if known"
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
[
{
"id": "cub_001",
"image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/4/45/Eopsaltria_australis_-_Mogo_Campground.jpg/1200px-Eopsaltria_australis_-_Mogo_Campground.jpg",
"context": "Identify the bird species and annotate visible attributes. Note any distinctive features."
},
{
"id": "cub_002",
"image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/6/6e/House_sparrow04.jpg/1200px-House_sparrow04.jpg",
"context": "Classify this bird and mark all visible body part attributes."
}
]
// ... and 1 more itemsGet This Design
Clone or download from the repository
Quick start:
git clone https://github.com/davidjurgens/potato-showcase.git cd potato-showcase/image/classification/cub-200-birds potato start config.yaml
Dataset & paper
Wah et al., 2011
Citation (BibTeX)
@techreport{wah2011caltech,
title = "The {Caltech-UCSD} Birds-200-2011 Dataset",
author = "Wah, Catherine and Branson, Steve and Welinder, Peter and Perona, Pietro and Belongie, Serge",
institution = "California Institute of Technology",
number = "CNS-TR-2011-001",
year = "2011",
url = "https://authors.library.caltech.edu/records/cvm3y-5hh21"
}Details
Annotation Types
Domain
Use Cases
Tags
Found an issue or want to improve this design?
Open an IssueRelated Designs
FLAIR: French Land Cover from Aerospace Imagery
Land use and land cover classification from high-resolution aerial imagery. Annotators classify the primary land use category of aerial image patches and identify any secondary land uses present. Based on the FLAIR dataset from the French National Institute of Geographic and Forest Information (IGN).
iWildCam Wildlife Detection & Classification
Camera trap image classification for wildlife monitoring (Beery et al., CVPR 2019). Classify wildlife species from camera trap images across diverse ecosystems worldwide.
KITTI Road Object Detection
Autonomous driving benchmark for object detection (Geiger et al., CVPR 2012). Annotate vehicles, pedestrians, and cyclists with 3D bounding boxes and occlusion/truncation labels.