Abstract

In the forensic science, fingerprint features play an important role in providing some clues about the criminal cases and identifying the suspect's identity. However, the latent fingerprint images from criminal scenes are often interfered by Gaussian noise, which leads to the blurring of the latent fingerprint images and affects the subsequent fingerprint identification. In recent years, the partial differential equation (PDE) method has made some progress in the field of image processing. In this paper, the application of the PDE method based on total variation is studied in latent fingerprint image enhancement processing, especially in the process of processing Gaussian noise in fingerprint images. In order to evaluate the performance of the PDE method in removing Gaussian noise in fingerprint images, Gaussian noise with different intensity is added to two different latent fingerprint images in the experiment in this paper. The experimental results of the PDE method are compared with the results of the neighborhood average method by peak signal-to-noise ratio and structural similarity, respectively. By comparison, it is discovered that when the intensity of Gaussian noise is low, the effect of the PDE method is almost the same as the effect of the neighborhood average method in processing Gaussian noise in fingerprint images. But with the increase of Gaussian noise intensity, the PDE method is better than the neighborhood average method in processing Gaussian noise of fingerprint images. The weight in the PDE method has a great influence on the effect of processing Gaussian noise in fingerprint images. The optimal weight is not fixed, which depends on the intensity of Gaussian noise and the fingerprint image itself.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call