Dataset of scattered images using noncoherent light under varying diffusion conditions and projected patterns

The data was acquired using a custom, low-cost optical setup built around a Raspberry Pi, which projects structured patterns through various optical diffusers and captures the resulting scattered light with a digital camera. All images are stored in PNG format with a resolution of 256×256 pixels and 8-bit depth.

Data collection was performed under different scattering conditions, achieved by employing multiple diffusers with varying physical thicknesses and scattering coefficients. The projected patterns comprise digits, geometric figures, and textures, providing a diverse set of visual features. Each dataset entry thus offers a direct mapping between a scattered observation and the original undistorted pattern, supporting research on deep learning-based image restoration and scattering compensation.