Somayeh Ebrahimkhani

How Do Medical MLLMs Fail? A Study on Visual Grounding in Medical Images

Guimeng Liu, Tianze Yu, Somayeh Ebrahimkhani, Lin Zhi Zheng Shawn et autres

Generalist multimodal large language models (MLLMs) have achieved impressive performance across a wide range of vision-language tasks. However, their performance on medical tasks, particularly in zero-shot settings where generalization is critical, remains suboptimal. A key research gap is the limited understanding of …

Brain‐FM: A Multimodal Foundation Model for Visual Question Answering in Brain Health Diagnostics

Somayeh Ebrahimkhani, Guimeng Liu, Tianze Yu, Xuling Lin et autres

BACKGROUND: Despite advancements in AI for diagnosing neurodegenerative diseases (NDs) like Alzheimer's (AD), existing models are task-specific and lack diagnostic reasoning, limiting their ability to interpret complex clinical and imaging data. Leveraging recent progress in multimodal foundation models (FMs) for healthcare, we …

sg (code pays fourni par la source)

Vision Transformer Neural Architecture Search for Out-of-Distribution Generalization: Benchmark and Insights

Sy-Tuyen Ho, Tuan Van Vo, Somayeh Ebrahimkhani, Ngai‐Man Cheung

While ViTs have achieved across machine learning tasks, deploying them in real-world scenarios faces a critical challenge: generalizing under OoD shifts. A crucial research gap exists in understanding how to design ViT architectures, both manually and automatically, for better OoD generalization. To …

Vision Transformer Neural Architecture Search for Out-of-Distribution Generalization: Benchmark and Insights

Sy-Tuyen Ho, Tuan Van Vo, Somayeh Ebrahimkhani, Ngai‐Man Cheung

Automated segmentation of knee articular cartilage: Joint deep and hand-crafted learning-based framework using diffeomorphic mapping

Somayeh Ebrahimkhani, Anuja Thimali Dharmaratne, Mohamed Hisham Jaward, Yuanyuan Wang et autres

my, au (code pays fourni par la source)

Evaluation of the GJB2 and GJB6 Polymorphisms with Autosomal Recessive Nonsyndromic Hearing Loss in Iranian Population.

Somayeh Ebrahimkhani, Golnaz Asaadi Tehrani

INTRODUCTION: (encoding connexin 32) genes, respectively. MATERIALS AND METHODS: In this study, 32 blood samples were obtained from Iranian patients with HL defect and 32 normal blood samples were prepared. After genomic deoxyribonucleic acid extraction, genotyping in rs7333214 and rs7329857 polymorphisms was …

ir (code pays fourni par la source)

A review on segmentation of knee articular cartilage: from conventional methods towards deep learning

Somayeh Ebrahimkhani, Mohamed Hisham Jaward, Flavia Cicuttini, Anuja Thimali Dharmaratne et autres

my, au, gb (code pays fourni par la source)