My research.

Object detection
under occlusion.

OCCLUDED OBJECT DETECTION FOR TRASH SORTING
Collaborating with Dr. Tae-Hoon Kim

Benchmarking YOLO11 and YOLO26 for object detection in cluttered, partially occluded scenes. I measure how detection changes as more of an object is hidden, using trash sorting as the test application.

I compare recall and average precision across object sizes and occlusion conditions to identify missed detections.

I also fine tune YOLO11n on occluded objects and compare detection of the target object with detection across the full image.

Medical image
generation.

QUANTUM ASSISTED MEDICAL IMAGE GENERATION
Mentored by Dr. Tae-Hoon Kim

I investigate models that combine quantum and classical components to generate synthetic medical images. I compare them with classical models to assess image quality and resource use.

My work includes designing the architecture, training the models, and comparing their computational requirements.

Depression detection
from speech and text.

DEPRESSION DETECTION USING SPEECH AND TEXT
Mentored by Dr. AmirHossein MajidiRad

I combined BERT text embeddings with MFCC audio features in a model for depression detection.

My combined model achieved an F1 score of 0.70, above the text baseline in my evaluation. I also studied model efficiency to reduce inference requirements.

Manuscript in preparation for arXiv submission · August 2026.