<p>Latent fingerprints are partial or incomplete fingerprints typically found at crime scenes. This enhancement is a critical step in identifying individuals from fingerprints left at crime scenes. It involves applying various image processing algorithms to improve the quality and contrast of the fingerprint patterns, making them more distinguishable from the background. Many deep learning and machine learning techniques have been applied to this problem due to their ability to automatically learn complex patterns and features from latent fingerprint images. However, accurate matching and detection with minimal time consumption remain major challenges. To enhance the latent fingerprint image, a novel Laplace Kernelized Piecewise Regression-based Progressive Generative Adversarial Network (LKPR-PGAN) is employed in five different processes, namely image acquisition, preprocessing, region of interest (ROI) segmentation, minutiae feature extraction, and matching. First, latent fingerprint images are collected from the dataset during the image acquisition phase. The images then undergo preprocessing to remove noise using a Laplace kernelized enhanced frost filtering technique. Following this, ROI segmentation is performed to extract the fingerprint image using the Rand index diagonal proximity method. Next, the Progressive Generative Adversarial Network (GAN) is employed for minutiae feature extraction and matching. Piecewise regression is used to extract different minutiae features from the image. The Hamann similarity coefficient is used for minutiae feature matching with the ground truth. The accurate matching results significantly improve latent fingerprint identification performance. Experimental results are described that the LKPR-PGAN method achieves higher accuracy in latent fingerprint identification with minimal time consumption compared to existing methods.</p>