Jun Ho Oh

Title: Jun Ho Oh: Inventor in Semiconductor Equipment and Machine Learning

Introduction

Jun Ho Oh is an inventor based in Gyeonggi-do, South Korea. His patent work focuses on image-based inspection and machine-learning methods for improving the accuracy of defect detection and prediction. He is associated with Semes Co., Ltd. and holds 2 patents.

Latest Patents

Oh’s listed patents address the creation and use of image datasets for inspection and model training.

- **Nozzle Inspection Unit and Substrate Treatment Apparatus Including the Same.** This invention describes a nozzle inspection unit designed to generate defect data to improve the detection of defective nozzles. It collects nozzle-related images, classifies them, combines images of normal and defective nozzles, and trains on the merged data to generate additional image data. The unit can be incorporated into a substrate treatment apparatus.

- **Method and Apparatus for Generating an Adaptive Training Image Dataset.** This invention presents a method for improving prediction accuracy by updating image labels based on a model’s predictions. It trains an initial model on a first image dataset, evaluates images in a second dataset, and revises labels when the assigned class differs from the predicted class. The updated datasets are then used to train a second model.

Career Highlights

Oh’s listed inventions connect semiconductor substrate processing with data-driven inspection. His work describes ways to expand and refine training data, supporting the development of models intended to identify defects and improve classification performance. He is associated with Semes Co., Ltd., a company involved in semiconductor manufacturing equipment.

Collaborations

The provided information lists Ji Hoon Yoo and Kwang Sup Kim as Oh’s coworkers. Their names are associated with his professional context, though specific roles in the listed patents are not provided.

Conclusion

Jun Ho Oh’s patent portfolio highlights image-data generation and adaptive model training for nozzle inspection and substrate treatment, reflecting an emphasis on improving automated defect detection.