Last Leaf
Narrative-based AI chatbot intervention for palliative care.
CuraX · Clinical research line
Research
LIMEN studies how AI reshapes human judgment, agency, and responsibility — and how systems can preserve the boundaries where humans must still judge.
Current LIMEN work is shown separately from ongoing clinical research and earlier work.
01 · Current Direction
Judgment Boundary Design
Research on human judgment in AI-mediated environments: when assistance becomes substitution, where human judgment must remain, and how systems can be designed to preserve it.
Judgment Displacement
How AI changes the role, location, or exercise of human judgment.
Judgment Boundary
Where human judgment must remain rather than becoming nominal approval.
Judgment-Preserving Design
How AI-mediated environments can preserve meaningful human judgment.
Healthcare is LIMEN’s first empirical testbed.
02 · Ongoing Clinical Research
Palliative & Supportive Care
Ongoing clinical research on narrative, communication, and decision-making in serious illness and palliative care.
03 · Earlier Work
Research since 2020
Work preceding the current LIMEN direction, preserved as distinct research lineages rather than retrospectively reframed.
Palliative & Supportive Care
Clinical and narrative research on dignity, life review, bereavement, and family experience in serious illness.
Life Review Therapy
Effects of life review therapy on emotional well-being in patients near the end of life.
Publication
A qualitative study of family centrality in modified dignity therapy experiences among Korean patients with advanced cancer: KCSG-PC22-10. Journal of Korean Medical Science, 2026.
Related published work
Effects of online group art therapy on psychological distress and quality of life after family bereavement: In COVID-19 pandemic. The Arts in Psychotherapy, 2023.
Dignity therapy for effective palliative care: a literature review. Kosin Medical Journal, 2022.
AI-based Emotional Assessment
Image-based emotional assessment, deep learning, and mental-health screening from structured drawings and visual features.
Depression & Anxiety Prediction Models
Development of deep learning-based depression and anxiety prediction models.
Publications
Machine learning-based depression screening model from pre-drawn structured mandala colorings. Current Psychology, 2026 · ePub August 2026.
A hybrid CNN-machine learning model for anxiety screening with mandala coloring patterns: feature extraction and classification performance. Acta Psychologica, 2026 · ePub July 2026.
Patents
Deep learning-based depression screening method and electronic device using mandala coloring analysis · Patent pending.
Deep learning-based anxiety screening method and electronic device using mandala coloring analysis · Patent pending.
PITR Multimodal Emotional Analysis
Multimodal emotional analysis combining object recognition and natural language processing in PITR using deep learning.
Publication
Object detection based on deep learning method in a person in the rain image. Journal of Korean Institute of Next Generation Computing, 2022.
Patents
Deep learning-based psychological analysis method using PITR · Patent pending.
Deep learning-based object recognition method for drawings and computer-readable recording medium storing the program · Korean Patent No. 10-2963703 · Registered May 7, 2026.
Medical Education
Publication
Issues with using part-task models in clinical skills assessment and preparation. Korean Journal of Medical Education, 2025.