Dewinda Julianensi Rumala, Peter van Ooijen, Reza Fuad Rachmadi, Anggraini Dwi Sensusiati, and I Ketut Eddy Purnama
Publication Paper
The proposed model aims to provide more powerful and robust classification performance compared to single CNN models, even when data is limited or unevenly distributed. Our findings show that the Deep-Stacked CNN outperforms existing methods in the same domain, using fewer parameters and computations. Qualitative evaluation demonstrated that the model behaves similarly to young physicians who learn from each other to provide better assessments and judgments. This characteristic enables the model to contribute to more objective and accurate decisions, which can greatly assist physicians in diagnosing brain diseases faster and more accurately. Furthermore, the model is transferable to other tasks and domains and is not limited to the task of brain disease classification. This capability provides insights into its generalizability and potential applications.
Publication link: read here.
Free access to paper (read only): read here.
