Automation of Detection of Cervical Cancer Using Convolutional Neural Networks

Abstract

Classification of digital cervical images acquired during visual inspection with acetic acid (VIA) is an important step in automated image-based cervical cancer detection. Many algorithms have been developed for classification of cervical images based on extracting mathematical features and classifying these images. Deciding the suitability of a feature and learning the algorithm is a complex task. On the other hand, convolutional neural networks (CNNs) self-learn most suitable hierarchical features from the raw input image. In this paper, we demonstrate the feasibility of using a shallow layer CNN for classification of image patches of cervical images as cancerous or not cancerous. We used cervix images acquired after the application of 3%-5% acetic acid using an Android device in 102 women. Of these, 42 cervix images belonged in the VIA-positive category (pathologic) and 60 in the VIA-negative category (healthy controls). A total of 275 image patches of 15 × 15 pixels were manually extracted from VIA-positive areas, and we considered these patches as positive examples. Similarly, 409 image patches were extracted from VIA-negative areas and were labeled as VIA negative. These image patches were classified using a shallow layer CNN composed of a layer each of convolutional, rectified linear unit, pooling, and two fully connected layers. A classification accuracy of 100% is achieved using shallow CNN.

Feasibility of deep learning for predicting live birth from a blastocyst image in patients classified by age

“…We used 50 × 50 pixels for the images of blastocysts. Only 15 × 15 pixels are used to detect cervical cancer . In a colposcopy study, it was reported that the accuracy for images of 150 × 150 pixels was better than that for 32 × 32 or 300 × 300 pixels, although images of uterine cervical lesion, including white epithelium and punctuation, seemed to be more complicated than images of blastocysts.…”

“…Only 15 × 15 pixels are used to detect cervical cancer. 93 In a colposcopy study, 67 it was reported that the accuracy TA B L E 4 Coefficients of the logistic regression, y = 1/ (1 + Exp(β 0 +β 1 x)), showing the probability of live birth as a function of the confidence score, which is the AI-generated predicted probability of live birth obtained from an image of the blastocyst When the AI system we made is applied to clinical use, the confidence scores could be used to select better blastocysts among all blastocysts according to the value. However, it is recommended that the regression function, which was applied to the data distribution of the patients, as shown in Figure 4 and Table 4, should be used to estimate the probability.…”

Feasibility of predicting live birth by combining conventional embryo evaluation with artificial intelligence applied to a blastocyst image in patients classified by age

“…Such a large number of datasets for each age category will improve the value of AUC, sensitivity, specificity, and accuracy of the classifier with deep learning. It is also considered to investigate the image size, the appropriate number of training datasets, the appropriate timing after insemination to capture images, and the regularization values for further study to improve the accuracies and to avoid overfitting that is an error that occurs when a classifier is too fit to a limited set of data. It would be better to study other parameters, such as information of time lapses regarding their potential to predict live birth.…”

Application of deep learning to the classification of uterine cervical squamous epithelial lesion from colposcopy images

“…Image information is one of the parameters that need to be investigated. Only 15×15 pixels are used to detect cervical cancer (72). In a colposcopy study (61), it was reported that the accuracy for images of 150×150 pixels was better than those for 32×32 or 300×300 pixels.…”