Update README.md

README.md

@@ -7,7 +7,11 @@ license: mit
 This study presents a machine learning-based approach to estimate optical vegetation indices and biophysical variables (hereafter referred to as VIs) using synthetic aperture radar (SAR) and ancillary data for forest monitoring. 
 The best-performing models were Random Forest Regressor (RFR) for LAI and FAPAR and XGBoost (XGB) for EVI and NDVI. These models were trained on temporally and spatially aligned time series (TS) datasets, containing Sentinel-1 SAR data, Sentinel-2 multispectral data, DEM-based features and meteorological variables. It provides an accurate and timely alternative to optical-based VIs.
 
-These models are part of the paper Paluba, D., Le Saux, B., Sarti, F., Štych, P. (2025): Estimating vegetation indices and biophysical parameters for Central European temperate forests with Sentinel-1 SAR data and machine learning. Published in Big Earth Data
+These models are part of the paper 
+> Paluba, D., Le Saux, B., Sarti, F., Štych, P. (2025): Estimating vegetation indices and biophysical parameters for Central European temperate forests with Sentinel-1 SAR data and machine learning. Published in Big Earth Data
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6798c936ece6b7910c55d1e5/3rueSUVk9bOqsFy4fsD-7.png)
+Figure 1. Methodology used in the paper.
 
 ## Model Details
 ### Model Description
@@ -68,6 +72,9 @@ To implement this model:
 
 The training data is available from the [SAR-based-VIs GitHub repository](https://github.com/palubad/SAR-based-VIs).
 
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6798c936ece6b7910c55d1e5/V49cLxspCqdoN_aURaD_c.png)
+Figure 2. Used areas for training and validation (training and validation data are not differentiated in this figure)
+
 ### Training Procedure
 
 - Feature Selection: Using permutation feature importance analysis to identify key predictors.
@@ -86,6 +93,13 @@ The training data is available from the [SAR-based-VIs GitHub repository](https:
 
 #### Training Hyperparameters
 
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6798c936ece6b7910c55d1e5/OkAYKL7hrEPIsYW0z4CgA.png)
+Figure 3. Hyperparameter tuning for NDVI.
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6798c936ece6b7910c55d1e5/TMFkP0TrsPlC-ROXcWy1M.png)
+Table 1. Best hyperparameter combinations identified for RFR and XGB. Bolded results represent the
+best achieved results for the VI.
+
 For detailed information on hyperparameter optimization, performances, speeds, please see the article Paluba et al. (2025).
 
 ## Evaluation metrics
@@ -93,7 +107,10 @@ For detailed information on hyperparameter optimization, performances, speeds, p
 - Mean Absolute Error (MAE): Primary metric for accuracy.
 - Mean Squared Error (MSE): Secondary metric for accuracy.
 - R² Score: To assess correlation with Sentinel-2 VIs.
-- Transferability Test: Applied to different Central European forests.
+- Transferability Test: Applied to different Central European forests (1,294 healthy deciduous and 1,253 healthy coniferous areas, and 1,195 disturbed areas).
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6798c936ece6b7910c55d1e5/9Kvzn0XgLB0oL6tFduyUb.png)
+Figure 4. Areas used to test the transferability of the models in Central Europe.
 
 ### Results
 
@@ -105,10 +122,12 @@ Best models:
 - SAR-based VIs detected forest changes up to 4 days earlier than Sentinel-2 VIs, significantly improving change detection capabilities.
 - Adding DEM and meteorological features improved R² by 3-4%.
 
+Table 2. Best results for RFR and XGB for each VI. Bolded results represent the best achieved results for the VI. 
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6798c936ece6b7910c55d1e5/1iLMhWwGJo6bbeb9COcSm.png)
 
 ## Citation [optional]
 
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+> Paluba, D., Le Saux, B., Sarti, F., Štych, P. (2025): Estimating vegetation indices and biophysical parameters for Central European temperate forests with Sentinel-1 SAR data and machine learning. Published in Big Earth Data
 
 **BibTeX:**