---
title: LaDeco
emoji: 👀
colorFrom: gray
colorTo: blue
sdk: gradio
sdk_version: 5.31.0
app_file: app.py
pinned: false
short_description: 'LaDeco: A tool to analyze visual landscape elements'
---

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

# LaDeco - Landscape Environment Semantic Analysis Model

LaDeco is a tool that analyzes landscape images, performs semantic segmentation to identify different elements in the scene (sky, vegetation, buildings, etc.), and enables region-based color matching between images.

## Features

### Semantic Segmentation
- Analyzes landscape images and segments them into different semantic regions
- Provides area ratio analysis for each landscape element

### Region-Based Color Matching
- Matches colors between corresponding semantic regions of two images
- Shows visualization of which regions are being matched between images
- Offers multiple color matching algorithms:
  - **adain**: Adaptive Instance Normalization - Matches mean and standard deviation of colors
  - **mkl**: Monge-Kantorovich Linearization - Linear transformation of color statistics
  - **reinhard**: Reinhard color transfer - Simple statistical approach that matches mean and standard deviation
  - **mvgd**: Multi-Variate Gaussian Distribution - Uses color covariance matrices for more accurate matching
  - **hm**: Histogram Matching - Matches the full color distribution histograms
  - **hm-mvgd-hm**: Histogram + MVGD + Histogram compound method
  - **hm-mkl-hm**: Histogram + MKL + Histogram compound method

## Installation

1. Clone this repository
2. Create a virtual environment: `python3 -m venv .venv`
3. Activate the virtual environment: `source .venv/bin/activate`
4. Install requirements: `pip install -r requirements.txt`
5. Run the application: `python app.py`

## Usage

1. Upload two landscape images - the first will be the color reference, the second will be color-matched to the first
2. Choose a color matching method from the dropdown menu
3. Click "Start Analysis" to process the images
4. View the results in the Segmentation and Color Matching tabs
   - Segmentation tab shows the semantic segmentation and area ratios for both images
   - Color Matching tab shows the matched regions visualization and the color matching result

## Reference

Li-Chih Ho (2023), LaDeco: A Tool to Analyze Visual Landscape Elements, Ecological Informatics, vol. 78.
https://www.sciencedirect.com/science/article/pii/S1574954123003187