Isolation of Thyroid Cancer Cells in FNAB Images using Image Processing Techniques

Fine-Needle Aspiration Biopsy (FNAB) is a non-surgical method of determining benign or malignant state of the thyroid nodule. To achieve high accuracy in the determination of benign or malignant state, the screening process should be carried out with an acceptable sensitivity. It is also necessary to develop efficient automated methods to perform the segmentation of cancer cells from the FNAB microscopic images before classification. In this paper, we have used an image segmentation method to perform cell image segmentation in Fine Needle Aspiration Biopsy Cytological Images of thyroid nodules under noise conditions and varying staining conditions. The segmentation method combines the image processing techniques thresholding, distance transform, watershed transform and Gaussian low pass filter. The watershed transform separates the overlapped cell images and Gaussian low pass filter prevents over segmentation by removing the noise objects in the image. The segmentation method is successfully tested for Papillary carcinoma and Medullary carcinoma Cytological Images of thyroid nodules, showing promising results, encouraging future research work.