Add visualization utilities for model analysis and reporting

[kamalahasiniburra]

Summary

This PR adds a reusable visualization utilities module for the DeepLense project with tests.

What is included

visualization_utils.py

• plot_image_grid(): Display sample images in a grid with labels
• • plot_training_history(): Plot loss and accuracy curves over epochs
• • plot_confusion_matrix(): Generate heatmap confusion matrices with annotations
• • generate_classification_report(): Create formatted text reports with per-class metrics

test_visualization_utils.py

• 7 tests covering all visualization functions
• • Tests for file saving, figure generation, RGB images, normalization, and report formatting

Why this is useful

Standardized visualization across DeepLense sub-projects helps ensure consistent model analysis and makes it easier to compare results between different approaches. All 7 tests pass locally.