Zenab Adnan Bosheah

Ms. Zenab Adnan Bosheah

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 February 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ZBZenab Adnan Bosheah
Zenab Adnan Bosheah
VBVilmos Bilicki
Vilmos Bilicki
Rapid advances in text-to-image (T2I) generative models have significantly enhanced visual content creation. However, evaluating these models remains challenging, particularly when assessing their ability to handle complex textual content. The primary aim of this research is to develop a systematic evaluation framework for assessing T2I models’ capabilities in generating specialized content, with emphasis on measuring text rendering accuracy and identifying model limitations across diverse domains. The framework utilizes carefully crafted prompts that require precise formatting, semantic alignment, and compositional reasoning to evaluate model performance. Our evaluation methodology encompasses a comprehensive assessment across many critical domains: mathematical equations, chemical diagrams, programming code, flowcharts, multi-line text, and paragraphs, with each domain tested through specifically designed challenge sets. GPT-4 serves as an automated evaluator, assessing outputs based on key metrics such as text accuracy, readability, formatting consistency, visual design, contextual relevance, and error recovery. Weighted scores generated by GPT-4 are compared with human evaluations to measure alignment and reliability. The results reveal that current T2I models face significant challenges with tasks requiring structural precision and domain-specific accuracy. Notable difficulties include symbol alignment in equations, bond angles in chemical diagrams, syntactical correctness in code, and the generation of coherent multi-line text and paragraphs. This study advances our understanding of fundamental limitations in T2I model architectures while establishing a novel framework for the systematic evaluation of text rendering capabilities. Despite these limitations, the proposed benchmark provides a clear pathway for evaluating and tracking improvements in T2I models, establishing a standardized framework for assessing their ability to generate accurate and reliable structured content for specialized applications.

22 April 2026 in IEEE Access
Verified
Journal Article
Open access
ZBZenab Adnan Bosheah
Zenab Adnan Bosheah
DKDorottya Keresztes
Dorottya Keresztes
PLPéter Lődár
Péter Lődár
PGPéter Gaszner
Péter Gaszner
JSJúlia Szebényi
Júlia Szebényi
VBVilmos Bilicki
Vilmos Bilicki
This study presents a comprehensive approach to nail condition classification using multi-illumination imaging and uncertainty-aware deep learning.We developed a diagnostic system capturing nail images under ten lighting conditions (normal and polarized versions of white, UV, red, green, and blue light) and classifying them into three categories: normal, fungal infections, and other pathologies. Our Conv-NeXt-based neural network architecture demonstrated exceptional performance for fungal detection, achieving accuracies ranging from 93.63% to 96.31% across the various illumination modes. Multi-illumination analysis revealed distinctive signatures, with white light channels demonstrating highest importance (32–33% of attention weights) and red light showing specific utility for fungal detection (11.1% vs. 5–7% for other conditions). Spatial analysis through Gradient-weighted Class Activation Mapping (Grad-CAM) identified that models focus primarily on nail edges when detecting fungal infections, with quantitative grid analysis showing edge regions contributing approximately 3 times more to classification decisions than central regions. Bayesian ensembles provided reliable uncertainty: predictive entropy discriminates correct vs. incorrect decisions (AUROC 0.9429), with strong calibration for the fungus classifier (Brier 0.040, ECE 0.032). t-SNE visualization demonstrated clear feature space separation, with fungal infections showing the most cohesive clustering corresponding to their distinctive illumination-spatial patterns. These convergent findings support clinical examination optimization: prioritizing nail edges under white light, with supplementary red light for suspected fungal cases, represents a promising strategy for future clinical validation.