README.md

metadata

title: Mars Vision Leaderboard
emoji: 🚀
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.19.2
app_file: run.py
pinned: false

Mars Vision Leaderboard

A comprehensive leaderboard for evaluating computer vision models on Mars-specific datasets. This leaderboard tracks performance across multiple tasks including classification, object detection, and segmentation.

Overview

This leaderboard provides a standardized evaluation framework for computer vision models on Mars-specific datasets.

Tasks

1. Classification

   • DoMars16k - Surface Types
   • Mars Image - Content Analysis
   • Deep Mars - Deep Learning
   • Dusty vs Non-dusty - Dust Analysis

2. Object Detection

   • Robins & Hynek - Craters
   • Lagain - Surface Features
   • SPOC - Surface Properties
   • AI4MARS - Surface Analysis
   • MarsData - General Surface

3. Segmentation

   • S5Mars - Surface
   • Mars-Seg - Features
   • Martian Landslide
   • Martian Frost

Getting Started

1. Clone the repository:

   git clone https://huggingface.co/spaces/gremlin97/mars-vision-leaderboard
   cd mars-vision-leaderboard
   

2. Install dependencies using Poetry:

   poetry install
   

3. Run the leaderboard:

   # From the project root directory
   poetry run python run.py
   

The leaderboard will be accessible at http://localhost:7860 when running locally.

Features

• Interactive Gradio interface
• Filter models by task
• Compare performance across datasets
• Visualize results with plots
• Track best performing models
• Detailed results table

Contributing

To add your model's results to the leaderboard:

1. Fork this repository
2. Add your results to the appropriate data dictionary in app/data.py
3. Submit a pull request with your changes

Results Format

For each task, results should be added in the following format:

TASK_DATA = {
    "Model": ["Model1", "Model2", ...],
    "Dataset": ["Dataset1", "Dataset1", ...],
    "Metric1": [value1, value2, ...],
    "Metric2": [value1, value2, ...],
}

Project Structure

mars-vision-leaderboard/
├── app/
│   ├── __init__.py
│   ├── app.py          # Main Gradio interface
│   ├── data.py         # Dataset and model data
│   └── leaderboard.py  # Visualization functions
├── run.py              # Application entry point
├── pyproject.toml      # Poetry dependencies
└── README.md