This project represents a collaborative effort between the Better Future Institute (BFI) of San Antonio and Clemson University. The primary dataset consisted of 311 non-emergency service requests collected over a 14-month period, which was provided to the research team. This dataset was subsequently enriched with demographic and socio-economic data for San Antonio, sourced from the U.S. Census Bureau, to facilitate a comprehensive analysis.
This research implements Exact Diffusion Inversion via Coupled Transformation (EDICT) with the Gaussian Shading watermarking technique in diffusion models. The results show a slight improvement in the performance of Gaussian Shading. The implementation is tested on manipulated images after watermarking, and results in the table below shows better performance for most image manipulation methods, except for ColorJitter and Salt & Pepper Noise. For more details, please refer to our paper.
