Applied AI & Computer Vision

Hossein Shirali

Doctoral Researcher · Karlsruhe Institute of Technology

AI research into real systems.

I connect computer vision, scientific imaging, and robotics to turn images into useful knowledge—with biodiversity as a central application.

About Me

I am a doctoral researcher at Karlsruhe Institute of Technology (KIT), supervised by apl. Prof. Dr. Christian Pylatiuk and apl. Prof. Dr. Markus Reischl. My research connects computer vision, scientific imaging, and robotics for biodiversity research. I develop methods for recognizing insects and measuring their traits, and integrate these methods into imaging and analysis systems.

My background spans electronics, data science, and applied machine learning, including work on industrial maintenance and satellite imagery. From May to August 2026, I completed a research stay in Prof. David Rolnick’s group at Mila. I enjoy working across disciplines to turn research questions into software and systems that people can use.

Research Interests

Fine-grained recognition

Computer vision for distinguishing visually similar species and extracting specimen-level information.

Scientific imaging & robotics

Automated acquisition and integrated AI workflows that connect physical specimens with digital analysis.

Biological measurement

Image-derived estimates of morphology, volume, and biomass for ecological research.

Selected Publications

All publications

InsectMorphoAI interface showing a fly, anatomical segmentation, and volume estimation. — InsectMorphoAI software interface.

Ecological Informatics · 2026Published

InsectMorphoAI: A deep learning framework for automated insect morphometrics and biomass estimation with taxon-specific volumetric validation

Estimates insect length, volume, and biomass from 2D images, with taxon-specific volumetric validation.

Shirali, H., Ascenzi, A., Wührl, L., Beyer, N., Di Lorenzo, N., Vaccarella, E., Klug, N., Meier, R., Cerretti, P., & Pylatiuk, C.

96, 103854.

Earlier preprint (2025)

Entomoscope photomicroscope with camera, motorized vertical stage, and illuminated specimen platform. — Entomoscope 2.0 device photograph, Shirali et al.

bioRxiv · 2026Preprint

Entomoscope 2.0 and ENIMAS 2.0: An open-source, AI-integrated platform for rapid and affordable insect digitization

Combines an affordable photomicroscope with open-source software for automated insect imaging and AI analysis.

Shirali, H., Boese, N. E., Kramer, M., Wang, J., Klug, N., Mikut, R., Meier, R., & Pylatiuk, C.

Four specimen images illustrating orientation classification and model attention on fly anatomy. — Shirali et al. (2026), CC BY.

Systematic Entomology · 2026Published

Automated specimen triage for dark taxa: Deep learning enables orientation, sex identification and anatomical segmentation from robotic imaging

Uses deep learning to identify specimen orientation and sex and segment anatomy in images of morphologically challenging flies.

Shirali, H., Wührl, L., Lee, L., Klug, N., Meier, R., Pylatiuk, C., & Hartop, E.

51, e70039.

Twelve labelled Agrilus beetle specimens with different body shapes and colours; specimens are not to scale. — Fig. 1, Caruso et al. (2026), CC BY 4.0. Specimens not to scale.

NeoBiota · 2026Published

Image-based recognition using advanced neural networks can aid surveillance of Agrilus jewel beetles

Tests neural networks for image-based identification of Agrilus beetles to support forest pest surveillance.

Caruso, V.^*, Shirali, H.^* · Full author list

Caruso, V.^*, Shirali, H.^*, Bouget, C., Cerretti, P., Curletti, G., de Groot, M., Groznik, E., Gutowski, J. M., Pylatiuk, C., Plewa, R., Roques, A., Sallé, A., Sweeney, J., Van Rooyen, K., Wührl, L., & Rassati, D.

^* Equal contribution.

105, 319-336.

Systems & Projects

Explore all projects

Let’s connect.

I welcome conversations about research collaboration and applied AI opportunities—especially where computer vision meets real-world systems.

hossein.shirali@kit.edu