International Journal of Chemical and Biochemical Sciences (ISSN 2226-9614)[/vc_column_text][/vc_column][/vc_row]

VOLUME 25(19) (2024)

Generate medical cell images that can be controlled by MCGAN

Xu Tao^*, Rowell Hernandez

Batangas State University, The National Engineering University Batangas City, Philippines

Abstract

Due to hospital privacy and policy restrictions and scarcity of medical image samples for medical image development, there are few medical image databases available for deep learning training and all access to medical image databases is difficult. This study identifies a research endeavor to generate medical cell images that can be controlled by masking cell position and overlap information. The experiments concluded that the MCGAN algorithm largely improves the generation quality of generative adversarial networks and enriches the diversity of manually collected data.

Keywords: MCGAN, High-precision cell detection, Generative adversarial network model, Diversity of data

Full length article – PDF   *Corresponding Author, e-mail: 21-03030@g.batstate-u.edu.ph    Doi # https://doi.org/10.62877/62-IJCBS-24-25-19-62