Update README.md

Author: Sayed-Hossein-Hosseini

README.md

@@ -49,13 +49,13 @@ The model follows a specialized three-stage pipeline:
 ### 1. Learning Curves
 In this section, you can observe the stability of the model during training. Thanks to **L2 Regularization** and **Global Gradient Clipping**, the training and validation curves converge smoothly without extreme oscillations.
 
-![Training History](results/learning_curves.png)  
+![Training History](Results/learning_curves.png)  
 *Figure 1: Loss and Accuracy curves for Training vs. Validation sets.*
 
 ### 2. Model Confusion Matrix
 The Confusion Matrix below illustrates the model's performance across all 5 classes. Note the high precision in the 'N' and 'Q' classes, and the improved sensitivity in the 'F' and 'V' categories.
 
-![Confusion Matrix](results/confusion_matrix.png)  
+![Confusion Matrix](Results/confusion_matrix.png)  
 *Figure 2: Confusion Matrix showcasing the identification of minority classes despite morphological similarities.*
 
 ---