Deep Learning in Forensic Sketch Analysis

When images of suspects are not accessible, forensic sketch analysis plays a crucial role in criminal investigations by giving investigators a visual representation of the subject. These sketches have typically been manually matched to databases or mugshots, which is a laborious and prone to human error process. This study investigates the use of cutting-edge deep learning methods to improve and automate forensic sketch recognition. In order to automate and improve sketch-to-photo recognition using public datasets, this work investigates sophisticated deep-learning techniques such as autoencoders, GANs, and CNNs. PCA and t-SNE are examples of dimensionality reduction algorithms that enhance feature representation and visualization. The outcomes indicate how well these models work to increase matching accuracy, providing law enforcement with a quicker and more dependable tool and highlighting AI’s potential to enhance criminal investigations and public safety.