TÜSEB
2025–2026
Ongoing
Automatic Staging of Acute Kidney Injury (AKI) with Deep Learning
Akut Böbrek Hasarı (AKI) Evresinin Derin Öğrenme ile Otomatik Tespiti
Group A Urgent R&D Project
PI: Fatma Tuana Doğu · Team: Hülya Doğan
KTÜ · Research Group · Since 2023
Where the lab bench meets the neural network.
Innovative AI Applications in Pharmacology
We design artificial intelligence and data-science methods that accelerate traditional pharmacology — from sharper microscopy and 3D cell imaging to drug interactions and mechanism-of-action prediction.
2023FoundedTrabzon, Türkiye
21Publications11 journal articles
7Funded projectsTÜBİTAK · TÜSEB · BAP
13MembersKTÜ · Trabzon Ü.
6SCI-E articlesindexed in Web of Science
Mission
Pharmacology, accelerated by AI
Our group carries out research that finds new solutions by applying artificial intelligence and data science to develop and accelerate traditional pharmacology methods. The work is led by researchers with complementary expertise in artificial intelligence, deep learning, data analytics, image processing and pharmacology.
Multidisciplinary
AI engineers and pharmacologists in one team, from wet lab to GPU.
Open datasets
We build and share expert-annotated microscopy and signal datasets.
Rigorous & ethical
Reproducible pipelines, careful validation and data governance.
Research
What we work on
Six connected directions, from the microscope to the clinic.
Extended Depth of Focus
One sharp image from a focal stack
Shape from Focus & 3D Microscopy
Recovering 3D structure from focus
Cellular Image Analysis
Segmenting and classifying cells
Drug Interactions & Repurposing
Predicting how drugs interact
Mechanism of Action from Signals
Reading drug effects in contractility
Clinical Decision Support
From lab models to patient care
Projects
Projects
Funded projects
Research supported by TÜBİTAK, TÜSEB and KTÜ BAP.
TÜBİTAK
2025–2026
Ongoing
Optimal Imaging and Novel Deep Learning Approaches for 2D and 3D Cell Segmentation in Microscopic Systems
Mikroskobik Sistemlerde 2B ve 3B Hücre Segmentasyonu için Optimum Görüntüleme Sağlanması ve Yeni Derin Öğrenme Tabanlı Yaklaşımlarının Tasarlanması
3501 – National Young Researchers Career Development Program
PI: Hülya Doğan · Team: Feride Sena Sezen, Ramazan Özgür Doğan
TÜSEB
2024–2025
Completed
Novel Deep Learning Approaches for Automatic Prediction of the Mechanisms of Action of Agents Affecting Smooth Muscle Contractility
Düz Kas Kontraktilitesini Etkileyen Ajanların Etki Mekanizmalarının Otomatik Tahmini için Derin Öğrenme Tabanlı Yeni Yaklaşımların Geliştirilmesi
Group A Urgent R&D Project
PI: Feride Sena Sezen · Team: Hülya Doğan, Z. Kalha, İlyas Ay
Publications
All publications
Selected publications
An effective pyramid vision transformer-based extended depth of focus technique for optimally focused microscopic imaging
The Journal of Supercomputing, 82(12), 634
From Handcrafted Focus Measurement Operators to Deep Learning: A Comprehensive Review of Shape from Focus Strategies
Archives of Computational Methods in Engineering
Mechanism of Action of Pharmacological Agents: A New Deep Learning-based Prediction Model Using Smooth Muscle Contractility Signals
The 12th International Conference on Computer Technology Applications (ICCTA 2026) · Vienna, Austria
Swin transformer-based shape-from-focus technique for accurate 3D shape estimation
Optics and Lasers in Engineering, 193, 109108
Open science
Code & Data
Code & data
Code and curated datasets we share with the community.
Hugging Face · dataset
Gated access
PCL-PASI: Multi-Task Multi-Device Psoriasis Dataset
pharmaiktu/pcl-pasi
Multi-task, multi-device image dataset for psoriasis severity assessment in mice: erythema, thickness and scaling scores, captured with three cameras over an 11-day experiment.
GitHub · repository
contractility-moa
pharmaiktu/contractility-moa
Two-stage framework that turns 1D smooth-muscle contractility signals into Gramian Angular Field images and predicts the mechanism of action with a Swin Transformer. Code is released after acceptance.
Hugging Face · dataset
Gated access
PharmAI KTU Contractility Signals
pharmaiktu/contractility-signals
578 isometric contractility time series from isolated murine trachea and bladder smooth muscle under agents with distinct receptor-mediated mechanisms of action.
News
All news
Latest news
PVT-EDOF published in The Journal of Supercomputing
Our pyramid vision transformer approach to extended depth of focus produces optimally focused microscopic images.
Read moreTwo papers at ICCTA 2026 in Vienna
Mechanism-of-action prediction from smooth muscle contractility signals and a neuro-focus fusion framework.
Read moreNew multi-focus dataset presented at Smart Multimedia 2025, Paris
Deep learning-based image fusion with a novel multi-focus dataset for extended depth of focus in microscopy.
Read moreJoin PharmAI
Interested in AI for pharmacology? Students (BSc projects, MSc, PhD), visiting scholars and partner labs are welcome. Send a short bio and your research interests by e-mail.