Ha Trong Nguyen // bibliography
Publications in OCR, document understanding, and time series forecasting. Selected venues only — full list available on request.
This paper revisits long-term time series forecasting by re-evaluating five representative Transformer-based and linear-based models on five real-world datasets under a unified protocol.
Food waste often stems from inaccessible packaging information. We introduce VNProductKIE, a dataset of Vietnamese food packaging images with English/Vietnamese text, and a three-stage pipeline: YOLO11x detection, word segmentation, and VietOCR recognition. Achieved 98.85% word-level accuracy.