Publications

Under Review

Statistical Distributions of Voxel Grid Resolution Induced Surface Deviations in Marching Cube-Based 3D Reconstructed Regular Smooth Geometries

Avinash Kumar K M, Samarth S. Raut

Submitted to Computer-Aided Design

Characterises statistical distributions (generalised gamma identified as best-fit via MLE + AIC/BIC/CvM) of voxel-resolution-induced surface errors across 9 geometries.

Preprint

Influence of Geometry, Class Imbalance and Alignment on Reconstruction Accuracy — A Micro-CT Phantom-Based Evaluation

Avinash Kumar K M, Samarth S. Raut

arXiv:2602.07658 (2026)

Validates reconstruction pipeline on micro-CT printed phantoms; studies effect of class imbalance and alignment on Dice, Jaccard, and Hausdorff metrics.

🔗 View on arXiv

In Preparation

3D Bioprinting in Disease Modeling: Integrating Mechanical and Clinical Cues for Physiologically Relevant Systems

Prakash Jayabal, Karolinekersin Enoch, Avinash Kumar K M, Samarth S. Raut, Suneel Rallapalli, Vijay Singh Parihar, Minna Kellomäki, Murugan Ramalingam, Santosh Mathapati

Conference

Error Assessment in Image-Based 3D Reconstruction and 3D Printing

Avinash Kumar K M, Samarth S. Raut

30th Congress of European Society of Biomechanics (ESB 2025) · ETH Zurich, Switzerland

Evaluates geometric fidelity in the pipeline from image-based 3D reconstruction to physical 3D printing. Proposes metrics and comparison frameworks to identify error accumulation.

📄 View PDF

Conference

Transfer-Learning Based Multi-Class Areca Nut Image Classification

Avinash Kumar K M, Aishwary Singh, Rahul L. Maurya, D. Narasimha, Samarth S. Raut

2024 IEEE Conference on Engineering Informatics (ICEI)

Deep learning model leveraging transfer learning to classify areca nuts under varied lighting conditions using a conveyor-based imaging system.

📄 View PDF

Conference

Active Flow Control Using Synthetic Jets and Neural Network

Hemanth M, Avinash Kumar K M, Aashish V Bhat, Bhagyashree Kulkarni, Dr. T. N. Shridhar

16th Annual CFD Symposium · NAL-CSIR, Bangalore · August 2014

Flow control experiments using synthetic jets, enhanced by neural network models to predict and adapt control strategies. Validated through wind tunnel experiments.

📄 View PDF