A multi-view approach to cDNA micro-array analysis

Abstract

Microarray has emerged as a powerful technology that enables biologists to study thousands of genes simultaneously, therefore, to obtain a better understanding of the gene interaction and regulation mechanisms. This paper is concerned with improving the processes involved in the analysis of microarray image data. The main focus is to clarify an image's feature space in an unsupervised manner. In this paper, the Image Transformation Engine (ITE), combined with different filters, is investigated. The proposed methods are applied to a set of real-world cDNA images. The MatCNN toolbox is used during the segmentation process. Quantitative comparisons between different filters are carried out. It is shown that the CLD filter is the best one to be applied with the ITE.Keywords: micro-array image processing; ITE; image transformation engine; median filter; top-hat filter; linear complex diffusion; adaptive segmentation; computational biology; drug design.Reference to this paper should be made as follows: Zineddin, B., Wang, Z., Shi, Y., Li, Y., Du, M. and Liu, X. (2010) 'A multi-view approach to cDNA micro-array analysis', Int.

Image-Based Quantitative Analysis of Gold Immunochromatographic Strip via Cellular Neural Network Approach

“…For this purpose, an image quality measurement, known as the peak signal-to-noise ratio (PSNR) [32], is used and the rational is justified as follows.…”

“…The higher the PSNR value the more strongly the binary spot mask fits with the raw image surface. The PSNR is a much better measure of the suitability of a spot mask to the raw image area containing the spot as it is more resilient against large intensity ranges within the spot area compared to the MSE [32]. Fig.…”

“…In the past decade, CNNs have gained much attention with wide applications in image and signal processing, robotics and pattern recognition. It is remarkable that CNN allows for efficient VLSI (very-large-scale integration) implementation of analogue, array-computing structures [32]. Recently, the CNN algorithm has been applied for the GICS (Gold immunochromatographic strip) image segmentation via adaptively setting the threshold value I in Ref.…”

Deep Belief Networks for Quantitative Analysis of a Gold Immunochromatographic Strip

“…In this part, a well-defined quality criterion called the peak signal-to-noise ratio (PSNR) [36] is exploited to quantify the segmentation performance of DBN method for GICS images, and also verify the effectiveness of the test and control lines in this study. In general, the PSNR is viewed as a measurement of peak error.…”

“…As an important performance indicator, the PSNR describes the ratio of the signal's peak value over the magnitude of the background noise. Generally, we prefer a larger PSNR value since the binary spot mask in this way fits better with the raw image surface [36]. Fig.…”