Research Showcase

Research Showcase#

Published research that has used VisualTorch to visualize model architectures.

Paper

Venue (Year)

Deep learning-based high-information-content graph representation of early stage bacterial biofilms

npj Biofilms and Microbiomes (2026)

Energy-Efficient Epileptic Seizure Prediction Using Spiking Neural Networks

IEEE ISCAS (2026)

Failure Evaluation of Steel Plate Shear Walls in Multi-Storey Steel Buildings Under Seismic Excitation Using Convolutional Neural Networks

Materials, MDPI (2026)

Deep learning-based high-information-content graph representation of early stage bacterial biofilms#

Authors: Nersesyan, L. E., Boiko, D. A., Kurbanalieva, S., Dzhemileva, L. U., Kozlov, K. S., Ananikov, V. P. (2026) Venue: npj Biofilms and Microbiomes Link: https://www.nature.com/articles/s41522-026-00971-3

Models early-stage bacterial biofilms as interaction graphs (cells as vertices, predicted intercellular interactions as edges), combining Mask R-CNN for cell segmentation with a custom network (BINet) for interaction prediction - enabling classification of developmental stage and substrate type from image-derived graph features.

BINet architecture

Energy-Efficient Epileptic Seizure Prediction Using Spiking Neural Networks#

Authors: Brady, A., Moore-Hill, D., Khan, F., Daoud, H. (2026) Venue: IEEE International Symposium on Circuits and Systems (ISCAS) Link: https://ieeexplore.ieee.org/document/11562867

A patient-specific model, trained on the CHB-MIT scalp EEG dataset, that combines convolutional layers with Leaky Integrate-and-Fire spiking neurons and a recurrent network to detect pre-ictal (pre-seizure) brain states - the low energy consumption of spiking neurons targets power-constrained, on-device seizure prediction for wearable/IoT devices.

Spiking neural network + LSTM architecture

Failure Evaluation of Steel Plate Shear Walls in Multi-Storey Steel Buildings Under Seismic Excitation Using Convolutional Neural Networks#

Authors: Bonfini, P., Schetakis, N., Sukhnandan, J., Drosopoulos, G. A., Stavroulakis, G. E. (2026) Venue: Materials (MDPI) Link: https://www.mdpi.com/1996-1944/19/5/878

Trains a CNN on physics-based finite element simulations to predict equivalent plastic strain (failure distribution) on steel plate shear walls from building geometry and seismic intensity, for use in structural digital twins.

CNN architecture

Used VisualTorch in your own research? Open a pull request to add it here.

Contents