EUVIMEDEUVIMED entdecken
Crossref

Hybrid Graph Mamba Framework for Histopathological Image Classification

Dave Machnaim, Shawn Jaison, Naveen Sundar, Narmadha Naveen D., Sai Midhun

Journal of Trends in Computer Science and Smart Technology · 24.04.2026 · 8 (2) · 304-323

DOI: 10.36548/jtcsst.2026.2.006 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Trends in Computer Science and Smart Technology, online erschienen am 24.04.2026. Verfasst von Dave Machnaim und 4 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 35 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Digital pathology has already been shown to be useful in disease diagnosis; however, the computational power required to use high-resolution digital pathology is substantial, making it unrealistic in many clinical laboratories due to limitations in hardware capabilities. Accordingly, this paper presents a new architecture, MFLA-Graph-Mamba-NCA, which addresses these limitations. The main point of MFLA-Graph-Mamba-NCA is to develop a lightweight model for tissue classification while maintaining speed and optimal accuracy. Four complementary learning paradigms are incorporated into the architecture: a morphological frontend, a frequency-domain branch based on Functional Kernel Adaptive Networks (FunKAN), feature stabilization over time using Liquid Neural Ordinary Differential Equations, and global sequence modelling using a Graph-based Mamba module. The model was tested on a harmonized benchmark of 125,000 images, based on the NCT-CRC-HE-100K and LC25000 datasets, and divided into 12 different biological classes to expressly avoid cross-dataset lab leakage. It has been shown to provide 99.89%accuracy and a Macro: F1 score of 99.78%. These techniques were condensed to 1.46 million parameters, forming a hybridized model that can be deployed to standard clinical workstations.

Zitieren

APA

Machnaim, D., Jaison, S., Sundar, N., D., N. N., & Midhun, S. (2026). Hybrid Graph Mamba Framework for Histopathological Image Classification. Journal of Trends in Computer Science and Smart Technology, 8(2), 304-323. https://doi.org/10.36548/jtcsst.2026.2.006

Vancouver

Machnaim D, Jaison S, Sundar N, D. NN, Midhun S. Hybrid Graph Mamba Framework for Histopathological Image Classification. Journal of Trends in Computer Science and Smart Technology. 2026;8(2):304-323. doi:10.36548/jtcsst.2026.2.006

BibTeX
@article{Machnaim2026,
  author = {Machnaim, Dave and Jaison, Shawn and Sundar, Naveen and D., Narmadha Naveen and Midhun, Sai},
  title = {Hybrid Graph Mamba Framework for Histopathological Image Classification},
  journal = {Journal of Trends in Computer Science and Smart Technology},
  year = {2026},
  volume = {8},
  number = {2},
  pages = {304--323},
  doi = {10.36548/jtcsst.2026.2.006}
}
Kostenlos starten →