A Deep Learning Framework for Accurate Classification of Ovarian Cancer
Abstract: Ovarian cancer is one among the deadliest forms of gynecological cancer, primarily due to its late diagnosis and the lack of reliable early screening methods. Recent advancements in deep learning have shown a great promise in medical image analysis, offering automatic approaches for cancer detection. This study investigates the performance of DenseNet and InceptionV3 in classifying ovarian cancer subtypes using histopathological images. The models were trained and tested on a relatively small dataset, employing various preprocessing and augmentation techniques to enhance performance. The findings of the study suggest that deep learning models can effectively classify ovarian cancer subtypes with good accuracy, even when data is limited. Among the models evaluated, DenseNet achieved and reached the highest classification with accuracy of 92%, demonstrating its potential as a suitable model for ovarian cancer diagnosis. These results underscore the need for the further optimization of deep learning frameworks to improve early detection and enhance clinical decision-making in ovarian cancer treatment.
Keywords: Ovarian Cancer, Deep Learning, DenseNet, InceptionV3, Histopathology, Medical Image Classification
How to Cite?: Abhilash T P, Pranathi P, "A Deep Learning Framework for Accurate Classification of Ovarian Cancer", Volume 14 Issue 5, May 2025, International Journal of Science and Research (IJSR), Pages: 624-627, https://www.ijsr.net/getabstract.php?paperid=SR25509153559, DOI: https://dx.doi.org/10.21275/SR25509153559