Akshatha Mohan

St. Louis, Missouri

I am a PhD student in Imaging Science at Washington University in St. Louis, where I work at the intersection of imaging science and machine learning. My background is in computer engineering (B.E. and M.S.), and my research interest is in building methods that turn complex imaging data into interpretable, clinically and scientifically meaningful measurements.

Broadly, I am interested in multimodal imaging, texture and statistical image analysis, explainable AI, and robust evaluation of imaging models. My work has spanned medical imaging (lung CT analysis, cellular morphology after nanoparticle exposure), computer vision (texture-aware deep learning, object detection), and remote sensing, with a recurring theme of asking not just whether a model works, but why it works and how we should measure it.

Previously, I earned my M.S. in Electrical and Computer Engineering at Texas A&M University, where I worked in the Advanced Vision and Learning Lab with Dr. Joshua Peeples, and my B.E. in Electronics and Communication at Bangalore Institute of Technology. I have also worked as an Android security software engineer and as a research intern in video processing and computer vision.

Feel free to look through my publications and projects, or reach out by email.

news

Aug 01, 2025	Started my PhD in Imaging Science at Washington University in St. Louis. 🎓
Feb 15, 2025	Our paper on texture analysis of lung cell morphology after nanoparticle exposure was presented (oral) at SPIE Medical Imaging 2025.
Feb 01, 2025	Began a collaboration with Duke University’s Virtual Imaging Trials team, studying false positives in AI-based lung cancer detection on clinical and simulated CT.
Dec 01, 2024	Completed my M.S. in Electrical & Computer Engineering at Texas A&M University.
Jun 17, 2024	Lacunarity Pooling Layers for plant image classification appeared at the CVPR 2024 Vision for Agriculture Workshop.

selected publications

SPIE
Texture analysis of lung cell surface morphology after nanoparticle exposure

Akshatha Mohan, Trevor Jefferis, Christie M. Sayes, and 1 more author

In Medical Imaging 2025: Digital and Computational Pathology, 2025

Abs Bib

We quantify morphological imaging metrics such as lacunarity and fractal dimension of lung cell surfaces from SEM images, relating them to dendritic cell sensitization following nanoparticle exposure, and introduce a Quantization and Counting Operator (QCO) for statistical texture analysis.
@inproceedings{mohan2025texture, title = {Texture analysis of lung cell surface morphology after nanoparticle exposure}, author = {Mohan, Akshatha and Jefferis, Trevor and Sayes, Christie M. and Peeples, Joshua}, booktitle = {Medical Imaging 2025: Digital and Computational Pathology}, volume = {13413}, pages = {221--230}, year = {2025}, organization = {SPIE}, }
CVPRW
Lacunarity Pooling Layers for Plant Image Classification using Texture Analysis

Akshatha Mohan and Joshua Peeples

In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, Jun 2024

Abs Bib HTML PDF

We introduce lacunarity pooling layers that leverage texture information to improve image classification, achieving up to 98.07% accuracy on PlantVillage and 95.00% on LeavesTex1200 with minimal additional parameters.
@inproceedings{mohan2024lacunarity, title = {Lacunarity Pooling Layers for Plant Image Classification using Texture Analysis}, author = {Mohan, Akshatha and Peeples, Joshua}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, pages = {5384--5392}, year = {2024}, month = jun, }
IGARSS
Quantitative Analysis of Primary Attribution Explainable Artificial Intelligence Methods for Remote Sensing Image Classification

Akshatha Mohan and Joshua Peeples

In IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium, 2023

Abs DOI arXiv Bib HTML

We evaluate primary-attribution XAI methods (LIME, HiResCAM, GradSHAP, Saliency, Occlusion) across multispectral and SAR imagery using a six-category metric framework spanning faithfulness, robustness, and localization.
@inproceedings{mohan2023quantitative, title = {Quantitative Analysis of Primary Attribution Explainable Artificial Intelligence Methods for Remote Sensing Image Classification}, author = {Mohan, Akshatha and Peeples, Joshua}, booktitle = {IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium}, pages = {950--953}, year = {2023}, doi = {10.1109/IGARSS52108.2023.10281981}, }