Sungyeon Park

Title: Sungyeon Park and Object Recognition Model Training

Introduction

Sungyeon Park is an inventor based in Seongnam-si, South Korea, and is associated with Hanwha Vision Co., Ltd. His patent work addresses object recognition model training, with a focus on improving how a model handles objects it detects incorrectly.

Latest Patents

Park is named on 1 patent: “Method for Training an Object Recognition Model in a Computing Device.” The disclosed method designates an object of interest for recognition and generates objects of non-interest to use as training data. When the model detects an object incorrectly, it can be retrained by automatically treating that object as one of non-interest, without requiring the user to provide feedback. This approach is intended to make error handling more convenient and improve the model’s reliability. The disclosure identifies potential applications in artificial intelligence, drones, robotics, augmented and virtual reality devices, and technologies related to 5G services.

Career Highlights

Park’s listed patent connects his work with computer vision and machine learning, particularly the training and refinement of object recognition models. His professional affiliation is Hanwha Vision Co., Ltd.

Collaborations

Park’s listed coworkers include Hyunhak Shin and Changho Song. They are associated with the patent information provided for this article.

Conclusion

Sungyeon Park’s patent describes a way to improve object recognition by using automatically generated non-interest examples to address incorrect detections, with potential relevance across several emerging technologies.