Thesis: ViT-based Aggregator and Group Features in Multiple Instance Learning for Whole Slide Image Classification and Segmentation (應用於全切片影像分類與分割的 ViT 聚合器與基於組特徵的多實例學習方法)
Thesis: Multiple Player Tracking with 3D Position Projection and Spatial-Temporal Information in Multi-View Volleyball Videos (多視角排球影片中考慮三維位置對應與時空資訊的多球員追蹤)
Thesis: A Multi-Scale Thermal-to-Visible Face Synthesis System with Studies of High-Frequency Information
and Race Bias Issues (結合高頻資訊及種族偏差研究的熱影像到可見光多尺度人臉合成系統)
Thesis: Object Class Invariant Model
for Age Estimation and Gender Classification in Pose Variant
and Occluded Faces (物件類別不動量模型應用於不同姿勢與遮蔽下之人臉年齡估計及性別分類)