Integrating Pap Smear Imaging for Enhanced Deep Learning-Based Cervical Cancer Screening

Le résumé fourni par la source

Cervical cancer is a worldwide health concern. In short, the detection of cervical cancer through the Pap smear test will certainly contribute to saving a large number of lives. But the cumbersome process of checking the results of the Pap smear test by the expert only, based on human intuition, results in biased outcomes. A number of deep learning and Machine learning techniques developed but the efficacy is limited for Pap smear images. In short, a novel automatic system for the detection of abnormal cervical cells by incorporating deep learning with Focused Mask Scoring R-CNN is proposed in this research paper which incorporate a lightweight squeeze-andexcitation (SE) block into the mask head to recalibrate channelwise feature responses. In the experiment, a high accuracy of 97.05 % is achieved by the model. In addition, the average precision, recall and$\mathbf{F 1}$-score of$0.97,0.98$, and 0.97 are obtained through the use of the testing data, respectively. Further, the area under the ROC curve is found to be greater than 0.95 by using the novel automatic system introduced in this paper, the experimental output shows the model is performing well as compare to existing state of art techniques.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.