A Deep Learning Ensemble Method to Assist Cytopathologists in Pap Test Image Classification

Abstract

In recent years, deep learning methods have outperformed previous state-of-the-art machine learning techniques for several problems, including image classification. Classifying cells in Pap smear images is very challenging, and it is still of paramount importance for cytopathologists. The Pap test is a cervical cancer prevention test that tracks preneoplastic changes in cervical epithelial cells. Carrying out this exam is important in that early detection. It is directly related to a greater chance of curing or reducing the number of deaths caused by the disease. The analysis of Pap smears is exhaustive and repetitive, as it is performed manually by cytopathologists. Therefore, a tool that assists cytopathologists is needed. This work considers 10 deep convolutional neural networks and proposes an ensemble of the three best architectures to classify cervical cancer upon cell nuclei and reduce the professionals’ workload. The dataset used in the experiments is available in the Center for Recognition and Inspection of Cells (CRIC) Searchable Image Database. Considering the metrics of precision, recall, F1-score, accuracy, and sensitivity, the proposed ensemble improves previous methods shown in the literature for two- and three-class classification. We also introduce the six-class classification outcome.

Automated Detection and Classification of Cervical Cancer Using Pap Smear Microscopic Images: A Comprehensive Review and Future Perspectives

“…Transfer learning based snapshot ensemble method (TLSE) is proposed for the fine-grained cervical cells classification task on Herlev dataset for 7 class and achieves an accuracy of 65%. [82] Authors address the big gap between the two and seven class accuracy and advise to GAN (Generative Adversarial Networks) [83] for increasing the number of samples.…”

“…The objective of the model is to maximize the number of true positive results and to minimize the number of false negative results. Authors have used convolutional neural networks and EfficientNet networks [83] which are state-of-the-art architectures for ImageNet dataset classification. The main block of architecture considered in the work includes MobileNet block, InceptionNet block and EfficientNet block.…”

“…The main block of architecture considered in the work includes MobileNet block, InceptionNet block and EfficientNet block. [83] For classification purpose, 80% of the images are used to train the networks and 20% for testing and 20% of the images are considered from training set for validation. The proposed ensemble model with these architectures used for the study portray an accuracy of 97% in cervical cell classification.…”

“…Calculation Formula Precision [72][73][74][75][76][77][78][79][80][81][82][83] TP/(TP + FP) [72][73][74][75][76][77][78][79][80][81][82][83] Recall (Sensitivity) [72][73][74][75][76][77][78][79][80][81][82][83] TP/(TP + FN) [72][73][74][75][76][77][78][79][80][81][82][83] Specificity [72][73][74][75]<...>…”

“…Specificity is the proportion of people who test negative among all those who do not have the disease. [17,72,83] Accuracy is the measure of correctly expected observation to the total observations. [17,72,83] F1 score is the weighted average of precision and recall which conveys the balance between both.…”