Lin Zou

Lin Zou

Auto-generated profile, based on publicly available information

Lin Zou

Synapxe, Elementum, Singapore

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

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

Representative Publications

View All Publications
25 February 2022 in IEEE Transactions on Artificial Intelligence
Verified
Journal Article
Open access
LZLin Zou
Lin Zou
HGHan Leong Goh
Han Leong Goh
CLCharlene Jin Yee Liew
Charlene Jin Yee Liew
JQJessica Lishan Quah
Jessica Lishan Quah
GGGary Tianyu Gu
Gary Tianyu Gu
ATAndy Wee An Ta
Andy Wee An Ta
Since the onset of the COVID-19 pandemic in 2019, many clinical prognostic scoring tools have been proposed or developed to aid clinicians in the disposition and severity assessment of pneumonia. However, there is limited work that focuses on explaining techniques that are best suited for clinicians in their decision making. In this article, we present a new image explainability method named ensemble AI explainability (XAI), which is based on the SHAP and Grad-CAM++ methods. It provides a visual explanation for a deep learning prognostic model that predicts the mortality risk of community-acquired pneumonia and COVID-19 respiratory infected patients. In addition, we surveyed the existing literature and compiled prevailing quantitative and qualitative metrics to systematically review the efficacy of ensemble XAI, and to make comparisons with several state-of-the-art explainability methods (LIME, SHAP, saliency map, Grad-CAM, Grad-CAM++). Our quantitative experimental results have shown that ensemble XAI has a comparable absence impact (decision impact: 0.72, confident impact: 0.24). Our qualitative experiment, in which a panel of three radiologists were involved to evaluate the degree of concordance and trust in the algorithms, has showed that ensemble XAI has localization effectiveness (mean set accordance precision: 0.52, mean set accordance recall: 0.57, mean set ${{\bm{F}}_1}$ : 0.50, mean set IOU: 0.36) and is the most trusted method by the panel of radiologists (mean vote: 70.2%). Finally, the deep learning interpretation dashboard used for the radiologist panel voting will be made available to the community. Our code is available at https://github.com/IHIS-HealthInsights/Interpretation-Methods-Voting-dashboard .

21 November 2025 in IEEE Access
Verified
Journal Article
Open access
MGMillie Ming Rong Goh
Millie Ming Rong Goh
LZLin Zou
Lin Zou
YTYen Ee Tan
Yen Ee Tan
MTMei Gie Tan
Mei Gie Tan
HGHan Leong Goh
Han Leong Goh
TKTse Hsien Koh
Tse Hsien Koh
Fungal infections represent a growing global health threat, especially for individuals with compromised immune systems. Accurate and timely species-level identification is essential for effective treatment but is often hindered by the labor-intensive and expertise-dependent nature of manual microscopic examination. This study presents FungAI, a two-stage deep learning pipeline for the automated classification of 20 clinically relevant fungal species from microscopic images. The first stage leverages deep learning models, including CNNs, Vision Transformers (ViTs), and Mamba-based architectures, to establish a robust classification backbone. While these models achieved high accuracy for filamentous fungi, they exhibited limitations in differentiating morphologically similar yeast species. To address this, a second-stage classifier was introduced and trained on handcrafted morphological features, enabling fine-grained discrimination among yeast classes. This addition led to significant performance gains, particularly when tested on the external test set, with accuracy improvements of up to approximately 10 percentage points. The best-performing model - a FungAI Two-Stage NextViT–MambaVision ensemble - achieved 92% accuracy across both holdout and external test sets. The proposed framework demonstrates strong potential for clinical deployment, reducing diagnostic turnaround time from up to 28 days to just 2 days. These results underscore the value of combining diverse architectural approaches with domain-specific morphological insights, setting a new benchmark for AI-driven medical mycology.