Paper Information | SBC BOOKS

Language (*)

Title (*)

Keywords (*)

Abstract (*)

Oral Squamous Cell Carcinoma (OSCC) is a prevalent malignancy with high mortality due to late-stage diagnosis, driven by challenges in interpreting subtle histopathological image features. We introduce GA-Channel-Xception, a novel bioinformatics framework that integrates a Genetic Algorithm (GA) with the Xception convolutional neural network to enhance the detection of OSCC from histopathological images. Our approach leverages GA to optimize color channel selection across RGB, HSV, and LUV color spaces, identifying the most informative three-channel combination for classification. A tailored Xception model, trained on synthetic images derived from these channels, was evaluated on a publicly available OSCC histopathological dataset. The proposed method achieved an accuracy of 98.71\% and an F1-score of 96.77\%, outperforming conventional approaches. These results underscore the potential of integrating evolutionary algorithms with deep learning to improve automated histopathological analysis, offering a scalable tool for early OSCC diagnosis and precision oncology.

Pages (*)

File Link

Title

Keywords

Abstract

Oral Squamous Cell Carcinoma (OSCC) is a prevalent malignancy with high mortality due to late-stage diagnosis, driven by challenges in interpreting subtle histopathological image features. We introduce GA-Channel-Xception, a novel bioinformatics framework that integrates a Genetic Algorithm (GA) with the Xception convolutional neural network to enhance the detection of OSCC from histopathological images. Our approach leverages GA to optimize color channel selection across RGB, HSV, and LUV color spaces, identifying the most informative three-channel combination for classification. A tailored Xception model, trained on synthetic images derived from these channels, was evaluated on a publicly available OSCC histopathological dataset. The proposed method achieved an accuracy of 98.71\% and an F1-score of 96.77\%, outperforming conventional approaches. These results underscore the potential of integrating evolutionary algorithms with deep learning to improve automated histopathological analysis, offering a scalable tool for early OSCC diagnosis and precision oncology.

#	Name
1	Joao Otavio Diniz(joao.bandeira@ifma.edu.br)
2	Breno Tamanini(tamaninib@acad.ifma.edu.br)
3	André Alvino(andrealvino@acad.ifma.edu.br)
4	Luana Cruz(luanabcruz@ufca.br)
5	Josenildo da Silva(jcsilva@ifma.edu.br)
6	Omar Andres Cortes(omar@ifma.edu.br)
7	Daniel Gomes(daniellima@ifma.edu.br)
8	Antonio de Carvalho Filho(antoniooseas@ufpi.edu.br)

Paper Registration

Fill in your paper information

English Information

Authors

Reference

#	Reference
No reference registered