Color Cervix Image Segmentation

Cervical cancer is the second most common cancer among women worldwide. In developing countries, cervical cancer is the leading cause of death from cancer. About 370,000 new cases of cervical cancer occur worldwide, resulting

(a) Fundus image (left stereo (b) Manually segmented optic image) disk/cup by an ophthalmologist on the right stereo image

(d) DA-segmented optic disk (e) DA-segmented optic cup (f) DA-segmented blood vessels

(g) Final segmentation of optic (h) Final segmentation of (i) Final segmentation of blood disk optic cup vessels

Figure 6.18: DA segmentation on clinical retinal image.

(g) Final segmentation of optic (h) Final segmentation of (i) Final segmentation of blood disk optic cup vessels

Figure 6.18: DA segmentation on clinical retinal image.

around 230,000 deaths each year. Cervical cancer develops slowly and has a detectable and treatable precursor condition known as dysplasia. It can be prevented through screening at-risk women and treating women with precancerous and cancerous lesions. In many western countries, cervical cancer screening programs have reduced cervical cancer incidence and mortality by as much as 90%. Analysis and interpretation of cervix images are important in early detection

IOC MO »0 iW

Segmented blood vessels (left)

Segmented blood vessels (right)

30 100 ISO 200 2i0 300 »0 -ICO

Feature map (left)

Feature map (right)

Feature map (right)

Interpolated disparity map Disparity map

(a) Disparity map with old segmentation technique Figure 6.19: Disparity maps generated with and without DA feature extraction.

Segmented blood vessels (left) Feature map (left)

Segmented blood vessels (right) Feature map (right)

Interpolated disparity map Disparity map (b) New DA-based segmentation technique

(c) Disparity map obtained from DA blood (d) Disparity map obtained from general vessel segmentation edge detection for blood vessel segmentation

(c) Disparity map obtained from DA blood (d) Disparity map obtained from general vessel segmentation edge detection for blood vessel segmentation

of cervical lesions. Automated image analysis is helpful in providing quantative lesion description thus monitoring of chronic lesions so that the onset of cervical cancer can be treated effectively.

A cervix image is a magnified color photograph of the cervix (illustrated in Fig. 6.20(a)). The acetowhite lesion area below the opening is marked by a trained physician, serving as a reference to other segmented images resulting from the algorithms. This image is taken with a regular high-resolution color camera, thus the most prominent problems preceding segmentation of the acetowhite lesion area from the rest are the reduction or removal of the glare and non-uniform illumination. Figure 6.20(b) shows the segmentation without illumination correction. All algorithms fail to recognize the lesion close to the cervix opening because the area is darker than other parts of the lesion, and vice versa for the section in the lower part of the cervigram, where the normal area is falsely classified as lesion. The glare on the top left of the lesion is also mis-classified. Figures 6.20(c)-6.20(f) are the segmentation results from k-means, DA, and AFLC after illumination correction. The results are similar, with DA generating the closest partition to the manual segmentation.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook


Post a comment