PSO ICA with BRM for Image Enhancement
Shih‐Hsiung Lee, Chu‐Sing Yang · 2016
In the teletransmission of images or medical imaging, image reconstruction or enhancement (denoising) is a significant topic. We can consider it as a typical issue of the Blind Source Separation (BSS) which interfere with transmission and cause blurs in images. Denoising and reconstruction refer to the removal of the unknown signals which lead to interference from the signals we intend to receive. Independent component analysis (ICA) has prominent performance in this field. This paper proposed to apply particle swarm optimization (PSO) algorithm to conduct accelerated computing of the rate of convergence of demixing matrix in ICA. Besides, it used Borel Regular Measure (BRM) based on Infomax to put forward a simple algorithm which was potential to be realized on a circuit chip. Our results of experiment have proved that after the separation of noises, the PSNR of the images has an apparent good effect. In addition, in the era of the Internet of Things, image and video perceptive products are facing the noises generated in transmission or by hardware interference. This paper will serve as an effective solution.