Abstract:Having a classifier of cell types in a breast cancer microscopic image (BCMI), obtained with immunohistochemical staining, is required as part of a computer-aided system that counts the cancer cells in such BCMI. Such quantitation by cell counting is very useful in supporting decisions and planning of the medical treatment of breast cancer. This study proposes and evaluates features based on texture analysis by fractal dimension (FD), for the classification of histological structures in a BCMI into either cancer cells or non-cancer cells. The cancer cells include positive cells (PC) and negative cells (NC), while the normal cells comprise stromal cells (SC) and lymphocyte cells (LC). The FD feature values were calculated with the box-counting method from binarized images, obtained by automatic thresholding with Otsu's method of the grayscale images for various color channels. A total of 12 color channels from four color spaces (RGB, CIE-L*a*b*, HSV, and YCbCr) were investigated, and the FD feature values from them were used with decision tree classifiers. The BCMI data consisted of 1,400, 1,200, and 800 images with pixel resolutions 128 × 128, 192 × 192, and 256 × 256, respectively. The best cross-validated classification accuracy was 93.87%, for distinguishing between cancer and non-cancer cells, obtained using the Cr color channel with window size 256. The results indicate that the proposed algorithm, based on fractal dimension features extracted from a color channel, performs well in the automatic classification of the histology in a BCMI. This might support accurate automatic cell counting in a computer-assisted system for breast cancer diagnosis.

Segmentation and classification of biological cell images by a multifractal approach

Efficient Segmentation Method Using Quantised and Non‐linear CeNN for Breast Tumour Classification

Fractal analysis segmentation of medical images: Application to breast cancer mammography image

A Fuzzy Segmentation Method to Learn Classification of Mitosis

Segmentation of cervical cell nuclei in high-resolution microscopic images: A new algorithm and a web-based software framework

Cell Segmentation Using Ellipse Curve Segmentation and Classification

Search for temporal cell segmentation robustness in phase-contrast microscopy videos

A Hybrid Skin Lesions Segmentation Approach Based on Image Processing Methods

A watershed-based algorithm to segment and classify cells in fluorescence microscopy images

Multispectral Spatial Characterization: Application to Mitosis Detection in Breast Cancer Histopathology

Principal component-based image segmentation: a new approach to outline in vitro cell colonies

Medical image segmentation based on multifractal theory

A Practical Segmentation Method for Automated Screening of Cervical Cytology

Cell type classifiers for breast cancer microscopic images based on fractal dimension texture analysis of image color layers

CellSeg: a robust, pre-trained nucleus segmentation and pixel quantification software for highly multiplexed fluorescence images

Unsupervised high-throughput segmentation of cells and cell nuclei in quantitative phase images

Image Processing for Automatic Cell Nucleus Segmentation Using Super pixel and Clustering Methods on Histopathological Images

Cell Segmentation in Microscopy Imagery Using a Bag of Local Bayesian Classifiers

Segmentation and ABCD rule extraction for skin tumors classification

Efficient cell segmentation tool for confocal microscopy tissue images and quantitative evaluation of FISH signals

A Framework for Segmentation of Inhomogeneous Live Cell Images using Fractional Derivatives and Level Set Method