Bio-Photonics Dataset: Light Scattering analysis for AI-Based Tissue Imaging
Description
The dataset consists of 2000 grayscale images (256 × 256 pixels) generated from Monte Carlo simulations of photon propagation in biological tissues. Each image represents the intensity distribution of photons after multiple scattering and absorption events within a simulated homogeneous biological medium, with pixel brightness corresponding to photon concentration (brighter areas indicating higher photon density). The dataset encompasses four key optical parameters systematically varied to capture the full range of light-tissue interactions: wavelength (450, 532, 650, and 780 nm), scattering coefficient (0.5–2.5 cm⁻¹), anisotropy factor (0.7–0.99), and absorption coefficient (0.1–1.5 cm⁻¹). Each image is accompanied by metadata documenting the specific optical properties used in its generation, stored in a CSV file that includes image path, wavelength, scattering coefficient, anisotropy factor, and absorption coefficient for all 2000 images. The dataset is organized in a single repository containing a IMAGES/ directory with all PNG files and a META_DATA.csv file. All images are stored in a flat directory structure without subfolders, following the naming convention N.png where N ranges from 0 to 1999. The metadata file contains 2000 rows with complete documentation for every image, including relative file paths and all four optical parameters. The dataset has been verified for image integrity, resolution consistency, and metadata completeness, with no missing files or values.
Files
Institutions
- Charotar University of Science and TechnologyGujarat, Anand