Cervical Cancer arises in the Cervix. Cancer progresses slowly from normal stage to pre-cancerous stage, and hence it provides an opportunity to diagnose it and prevent it. Visual inspection of the pap smear is tedious and time-consuming. Commercial systems for automated analysis are costly which makes it unaffordable for most hospitals. The primary aim of this research work is to develop a computer-aided diagnosis tool to identify if the cervical cancer is Squamous Cell Carcinoma or Adenocarcinoma by segmenting the pap smear images. Image segmentation and boundary detection techniques are utilized to separate the nucleus from the cytoplasm. Automated cropping procedure is implemented to have one cell per image that allows extracting the cytoplasm features. The work extracts texture, shape and pixel value measurements to classify cancer into its respective categories. It also classifies Squamous Cell Carcinoma as Low-grade Squamous Intraepithelial Lesion (LSIL) or High-grade Squamous Intraepithelial Lesion (HSIL). It usually takes three weeks for the pathologists to generate the Pap smear results. The proposed work identifies the stage of cancer within a few seconds. It works for the images with different stains and images with the single cell and multiple cells. The proposed work achieved an accuracy of 87.5 percent when the results obtained are compared with the ground truth. The promising results show that it can be used as a diagnostic tool by the pathologists.
[1]
Jianwei Zhang,et al.
Feature Quantification and Abnormal Detection on Cervical Squamous Epithelial Cells
,
2015,
Comput. Math. Methods Medicine.
[2]
Dwi H. Widyantoro,et al.
Extraction and classification texture of inflammatory cells and nuclei in normal pap smear images
,
2015,
2015 4th International Conference on Instrumentation, Communications, Information Technology, and Biomedical Engineering (ICICI-BME).
[3]
G. Kemper,et al.
A Detection Method of Ectocervical Cell Nuclei for Pap test Images, Based on Adaptive Thresholds and Local Derivatives
,
2015,
MUE 2015.
[4]
Debdoot Sheet,et al.
Unsupervised Segmentation of Overlapping Cervical Cell Cytoplasm
,
2015,
ArXiv.
[5]
Bai Ying Lei,et al.
Accurate Segmentation of Cervical Cytoplasm and Nuclei Based on Multiscale Convolutional Network and Graph Partitioning
,
2015,
IEEE Transactions on Biomedical Engineering.
[6]
Yunhui Liu,et al.
Accurate Segmentation of Partially Overlapping Cervical Cells Based on Dynamic Sparse Contour Searching and GVF Snake Model
,
2015,
IEEE Journal of Biomedical and Health Informatics.
[7]
Eko Supriyanto,et al.
Automatic detection system of cervical cancer cells using color intensity classification
,
2011
.