Vahid Khodadadi - Department of Biomedical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
Morteza Behnam - Department of Electrical Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Isfahan, Iran.
Pooyan Alinaghi Hosseinabadi - Department of Electrical Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Isfahan, Iran

The correct and detailed segmentation of cells are very critical for cellular dynamic observation in biologic. The provided novel algorithm Cell Segmentation Comprehensive Algorithm (CSCA) is a comprehensive and full automatic without away type and number of cells limitation for segmentation based on a novel method by inverse wavelet transform. In our algorithm, first 3D image slices are preprocessing separately. Then every noise slice is decreased by passing off a BM3D filter with optimum parameters. So, after histogram equalization in every slice image, we decreased the image background by a new method that passing in parallel of different Gaussian filter and subtract two images filtered by Gaussian filter. Then in the next stage, the object edges are recognized by using Gradient Vector Flow (GVF) and morphological operators, then corresponding novel formula with image structure data is used for omitting the picture omissions. Finally, by applying 2D wavelet transform from each slice, every four components of 2D wavelet transform in a matrix are stored separately and after processing of all slices and saving in four 3D matrixes. Then it is converted to 3D image by using 3D inverse wavelet transform in four matrixes. Finally, we detect and segment all of the cells that are inside or interface of frames 3D as a visual frame. The given algorithm on data has been tested with accuracy rate of 97.22%.

Segmentation; GVF; BM3D; Inverse Wavelet Transform; Local OTSU; Structure Local