Detection of Specular Reflection and Segmentation of Cervix Region in Uterine Cervix Images for Cervical Cancer Screening

Kudva, Vidya; Prasad, Keerthana; Guruvare, Shyamal

doi:10.1016/j.irbm.2017.08.003

Cited by 44 publications

(21 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In [12], cervix segmentation serves as the first step for the cervix image registration workflow; similar to [11], a fuzzy C-means clustering approach is employed for cervix segmentation and is reported to achieve an average IoU score of 0.76. In [13], cervix images captured using cell-phone cameras are segmented through the combination of edge detector, red-color component filter, curvature filter, and thresholding. The researchers report an average Dice score of 0.9, on a small set of 151 images.…”

Section: Conventional Image Processing Techniquesmentioning

confidence: 99%

Cross-Dataset Evaluation of Deep Learning Networks for Uterine Cervix Segmentation

2020

View full text Add to dashboard Cite

Evidence from recent research shows that automatic visual evaluation (AVE) of photographic images of the uterine cervix using deep learning-based algorithms presents a viable solution for improving cervical cancer screening by visual inspection with acetic acid (VIA). However, a significant performance determinant in AVE is the photographic image quality. While this includes image sharpness and focus, an important criterion is the localization of the cervix region. Deep learning networks have been successfully applied for object localization and segmentation in images, providing impetus for studying their use for fine contour segmentation of the cervix. In this paper, we present an evaluation of two state-of-the-art deep learning-based object localization and segmentation methods, viz., Mask R-convolutional neural network (CNN) and MaskX R-CNN, for automatic cervix segmentation using three datasets. We carried out extensive experimental tests and algorithm comparisons on each individual dataset and across datasets, and achieved performance either notably higher than, or comparable to, that reported in the literature. The highest Dice and intersection-over-union (IoU) scores that we obtained using Mask R-CNN were 0.947 and 0.901, respectively.

show abstract

Section: Conventional Image Processing Techniquesmentioning

confidence: 99%

Cross-Dataset Evaluation of Deep Learning Networks for Uterine Cervix Segmentation

2020

View full text Add to dashboard Cite

show abstract

“…Specular reflection is often encountered in colposcopy images and interferes with the true change of grayscale intensities [21], [27]- [29]. A threshold method was therefore used.…”

Section: Image Segmentationmentioning

confidence: 99%

“…Nevertheless, one limitation of the VIA test is its dependence on operator's subjective definition of a lesion region [18], [19], which are based on main attributes of color, vascular patterns and lesion margins [20]. Several factors could interfere with the visual inspection, e.g., the mucus on the cervical surface or reflective spots [21], making CIN classification challenging for less experienced operators and the application of VIA test in regions lacking experienced doctors difficult. The recent advance of machine learning in medical imaging analysis has provided a promising solution.…”

Section: Introductionmentioning

confidence: 99%

A Fuzzy Reasoning Model for Cervical Intraepithelial Neoplasia Classification Using Temporal Grayscale Change and Textures of Cervical Images During Acetic Acid Tests

2019

View full text Add to dashboard Cite

A novel fuzzy reasoning model using temporal grayscale change and texture information of acetic acid test cervical images was developed to classify the patients at risk for Cervical Intraepithelial Neoplasia (CIN). Methods: Pre-and post-acetic acid test images were obtained from 505 patients (383 CIN negative and 122 CIN positive). An automatic image segmentation algorithm was implemented to extract the acetowhite region, followed by feature extraction of the: 1) temporal grayscale change from pre-to post-images; 2) texture coarseness; and 3) texture complexity of post-test images. The index of CIN was derived based on fuzzy reasoning incorporating expert knowledge. Logistic regression was built with cross validation to compare the classification performance between different combinations of features. Results: The averaged sensitivity was 80.8%, 80.9%, and 82.8% and specificity was 82.0%, 87.4%, and 86.2% for three individual features, respectively. The combination of all three features significantly improved the sensitivity (84.7%) without a significant loss of specificity (86.5%). The fuzzy reasoning model further improved the sensitivity significantly (sensitivity 85.9% and specificity 86.6%). Conclusion: Image texture features outperformed temporal grayscale change in specificity. Combining grayscale-and texture-based features could improve the overall classification performance, creating a synergistic effect. The proposed fuzzy reasoning model could further increase the sensitivity without requiring additional features. Significance: The proposed automatic CIN classification algorithm could provide a useful screening tool in the prevention of cervical cancers. INDEX TERMS Cervical cancer screening, acetic acid test, fuzzy reasoning, texture feature.

show abstract

“…We use a method explained in [31] for ROI detection. Different stages used by this method for ROI detection in cervix images are shown in Fig.…”

Section: Roi Detectionmentioning

confidence: 99%

Andriod Device-Based Cervical Cancer Screening for Resource-Poor Settings

2018

View full text Add to dashboard Cite

Visual inspection with acetic acid (VIA) is an effective, affordable and simple test for cervical cancer screening in resource-poor settings. But considerable expertise is needed to differentiate cancerous lesions from normal lesions, which is lacking in developing countries. Many studies have attempted automation of cervical cancer detection from cervix images acquired during the VIA process. These studies used images acquired through colposcopy or cervicography. However, colposcopy is expensive and hence is not feasible as a screening tool in resource-poor settings. Cervicography uses a digital camera to acquire cervix images which are subsequently sent to experts for evaluation. Hence, cervicography does not provide a real-time decision of whether the cervix is normal or not, during the VIA examination. In case the cervix is found to be abnormal, the patient may be referred to a hospital for further evaluation using Pap smear and/or biopsy. An android device with an inbuilt app to acquire images and provide instant results would be an obvious choice in resource-poor settings. In this paper, we propose an algorithm for analysis of cervix images acquired using an android device, which can be used for the development of decision support system to provide instant decision during cervical cancer screening. This algorithm offers an accuracy of 97.94%, a sensitivity of 99.05% and specificity of 97.16%.

show abstract

Detection of Specular Reflection and Segmentation of Cervix Region in Uterine Cervix Images for Cervical Cancer Screening

Cited by 44 publications

References 13 publications

Cross-Dataset Evaluation of Deep Learning Networks for Uterine Cervix Segmentation

Cross-Dataset Evaluation of Deep Learning Networks for Uterine Cervix Segmentation

A Fuzzy Reasoning Model for Cervical Intraepithelial Neoplasia Classification Using Temporal Grayscale Change and Textures of Cervical Images During Acetic Acid Tests

Andriod Device-Based Cervical Cancer Screening for Resource-Poor Settings

Contact Info

Product

Resources

About