SyniSCAT is a Python-based simulation pipeline designed to generate synthetic datasets for Interferometric Scattering (iSCAT) microscopy. It provides a framework for creating large-scale training data for deep learning models (such as Transformers) by simulating the optical scattering of non-spherical particles and complex background noise. Key Features: Scattering Approximation: Implements a "rigid sphere cluster" model to approximate the scattering patterns of complex particle shapes using superposition. Defect Modeling: Explicitly simulates lithography artifacts, including nano-hole irregularities, edge roughness, and "double-dipping" effects. Data Pipeline: Automates the generation of full video sequences paired with ground-truth segmentation masks. Michael A Khanna; LION LAB at Vanderbilt code repository: https://github.com/michaelakhanna/syniSCAT