This research enhances composite sketches of individuals suspected of involvement in criminal activity using a Pix2Pix CycleGAN model architecture. The study focuses on the lengthy process of matching composite sketches to corresponding mug shots or images from official databases. This is a common challenge faced by law enforcement agencies. This enhancement facilitates quicker case resolution and reduces the backlog of criminal cases. The evidence demonstrates the model’s effectiveness in improving sketches to more closely resemble realistic photographs. The study uses the CUHK (CUFS) Face Sketch Database, which contains 606 images along with their corresponding sketches. The primary objective of this work is to minimize the manual effort required in linking composite sketches with actual suspects by automating the sketch-to-suspect identification process. This study seeks to explore the use of artificial intelligence and adversarial networks in areas where they have not yet been applied. The model achieved a Peak Signal-to-Noise Ratio (PSNR) of 67.48 dB. Recent advancements have highlighted the powerful capabilities of GAN models, and the current work aims to further refine and advance these methods.