Banana Ripeness Stage Image Dataset for LADDER Model Training and Evaluation

This dataset provides real-world banana images captured over five consecutive days to document the natural progression of ripeness under controlled indoor lighting. Images were acquired using a Moto G40 Fusion Android smartphone at 1080 \u00d7 1080 pixel resolution with a uniform white background to minimize visual noise and emphasize texture, color, and surface variations. The dataset contains 1,115 raw JPEG images organized into five folders (Day 1 \u2013 Day 5), corresponding to greenish-yellow, yellow, ripe, and overripe stages. These samples were collected to support research on the LADDER (Latent Autoencoding and Denoising Diffusion for Enhanced Realism) framework and other agricultural computer-vision applications such as fruit classification, ripeness estimation, and image-quality enhancement. The dataset is intended for research and educational use to facilitate reproducibility and benchmarking in diffusion-based and deep-learning-based image-refinement studies.