CHART-Infographics: Chart and Infographic Analysis
Chart and infographic analysis with structured extraction. Annotators identify chart elements (axes, legends, data points, titles) with bounding boxes, classify chart types, and extract data values. Supports structured understanding of visual data representations.
File di configurazioneconfig.yaml
# CHART-Infographics: Chart and Infographic Analysis
# Based on Davila et al., ICPR 2024
# Paper: https://link.springer.com/chapter/10.1007/978-3-031-78312-8_5
#
# Chart and infographic understanding requires identifying structural elements
# (axes, legends, titles, data points) and extracting the underlying data.
# This task combines visual element localization with semantic understanding.
#
# Annotation Guidelines:
# 1. First classify the chart type
# 2. Draw bounding boxes around key structural elements:
# - Chart Title, Axis Labels, Legend, Data Points/Bars/Slices
# 3. Extract readable data values from the chart
# 4. Note any special features (annotations, trend lines, dual axes)
#
# Element Types for Bounding Boxes:
# - Title: The main title of the chart
# - X-Axis Label: Label text along the horizontal axis
# - Y-Axis Label: Label text along the vertical axis
# - Legend: The legend/key area
# - Data Region: The area containing actual data visualization
# - Annotation: Any text annotations or callouts on the chart
annotation_task_name: "CHART-Infographics: Chart Analysis"
task_dir: "."
data_files:
- sample-data.json
item_properties:
id_key: "id"
text_key: "image_url"
output_annotation_dir: "annotation_output/"
output_annotation_format: "json"
annotation_schemes:
# Step 1: Identify chart elements with bounding boxes
- annotation_type: image_annotation
name: chart_elements
description: "Draw bounding boxes around the key structural elements of the chart."
tools:
- bbox
labels:
- "Title"
- "X-Axis Label"
- "Y-Axis Label"
- "Legend"
- "Data Region"
- "Annotation"
# Step 2: Classify chart type
- annotation_type: radio
name: chart_type
description: "What type of chart or visualization is this?"
labels:
- "Bar"
- "Line"
- "Pie"
- "Scatter"
- "Area"
- "Heatmap"
- "Other"
keyboard_shortcuts:
"Bar": "1"
"Line": "2"
"Pie": "3"
"Scatter": "4"
"Area": "5"
"Heatmap": "6"
"Other": "7"
tooltips:
"Bar": "Bar chart (vertical or horizontal bars representing values)"
"Line": "Line chart (data points connected by lines showing trends)"
"Pie": "Pie or donut chart (circular chart showing proportions)"
"Scatter": "Scatter plot (individual data points plotted on x-y axes)"
"Area": "Area chart (filled region between line and axis)"
"Heatmap": "Heatmap or matrix visualization (color-coded grid)"
"Other": "Any other chart type not listed above (radar, treemap, Sankey, etc.)"
# Step 3: Chart complexity
- annotation_type: radio
name: chart_complexity
description: "How complex is this chart?"
labels:
- "Simple"
- "Moderate"
- "Complex"
tooltips:
"Simple": "Single data series, clear labels, no annotations (e.g., simple bar chart)"
"Moderate": "Multiple data series, legend present, some annotations"
"Complex": "Multiple chart types combined, dual axes, dense data, or infographic elements"
# Step 4: Data extraction
- annotation_type: text
name: extracted_data
description: "Extract the key data values from the chart. List the data as label: value pairs, one per line."
# Step 5: Additional notes
- annotation_type: text
name: notes
description: "Note any special features, issues with readability, or additional observations about the chart."
html_layout: |
<div style="margin-bottom: 10px; padding: 8px; background: #f0f4f8; border-radius: 4px;">
<strong>Chart Source:</strong> {{chart_source}}
</div>
<div style="margin-bottom: 8px; padding: 8px; background: #fafafa; border-radius: 4px;">
<strong>Description:</strong> {{description}}
</div>
<div style="text-align: center; margin-bottom: 15px; padding: 10px; background: #ffffff; border: 1px solid #e5e7eb; border-radius: 6px;">
<img src="{{image_url}}" style="max-width: 100%; max-height: 600px;" />
</div>
allow_all_users: true
instances_per_annotator: 50
annotation_per_instance: 2
allow_skip: true
skip_reason_required: false
Dati di esempiosample-data.json
[
{
"id": "chart_001",
"image_url": "https://example.com/chartinfo/bar_chart_revenue_001.png",
"chart_source": "Annual Report 2023",
"description": "Grouped bar chart showing quarterly revenue for three product lines over fiscal year 2023."
},
{
"id": "chart_002",
"image_url": "https://example.com/chartinfo/line_chart_temperature_002.png",
"chart_source": "Climate Research Paper",
"description": "Multi-line chart displaying average monthly temperatures across five major cities from January to December."
}
]
// ... and 8 more itemsOttieni questo design
Clone or download from the repository
Avvio rapido:
git clone https://github.com/davidjurgens/potato-showcase.git cd potato-showcase/multimodal/chartinfo-chart-analysis potato start config.yaml
Dettagli
Tipi di annotazione
Dominio
Casi d'uso
Tag
Hai trovato un problema o vuoi migliorare questo design?
Apri un problemaDesign correlati
DocBank Document Layout Detection
Document layout analysis benchmark (Li et al., COLING 2020). Detect and classify document elements including titles, abstracts, paragraphs, figures, tables, and captions.
DocLayNet Document Layout Analysis
Document layout analysis with bounding box annotations. Annotators draw bounding boxes around layout elements (text blocks, tables, figures, headers, footers, lists) in document page images.
MOCHEG: Multi-modal Multi-hop Fact-Checking with Explanations
Multimodal fact-checking requiring reasoning over both text and images. Annotators verify claims using multimodal evidence, identify relevant spans in text evidence, and provide explanations for their verdicts. Supports multi-hop reasoning across modalities.