Neighborhood matrix: A new idea in matching of two dimensional gel images

Document Type: Original Research Papers



Automated data analysis and pattern recognition techniques are the requirements of biological and proteomics
research studies. The analysis of proteins consists of some stages among which the analysis of two dimensional
electrophoresis (2-DE) images is crucial. The aim of image capturing is to generate a Photostat that can be used in
future works such as image comparison. The researchers introduced a new method for matching two 2-DE gel
images. In this method, a neighborhood circular region is defined to obtain information about spots’ neighbors. In
the present paper, the information obtained by this region is reordered into a matrix as a descriptor of the neighbors
of each spot. The matrix is then used in matching the spots between two images. All conducted tests to evaluate the
method’s performance showed the power of the method in spot matching, even when the number of candidate
matching spots in the second images increased. The proposed method provides a robust automatic comparison idea
in gel images matching. Despite its low speed, its accuracy is excellent. The Novelty of the present study is the use
of matrices as neighborhood descriptor. This idea is applicable in any other similar domain.