Eun Soo Im

Title: Eun Soo Im and Multi-Camera Object Tracking

Introduction

Eun Soo Im is an inventor associated with Suwon-si, South Korea, and Superb AI Co., Ltd. His patent work addresses the challenge of tracking multiple objects across multiple cameras by using a temporal-spatial graph neural network.

Latest Patents

The supplied information lists 1 patent for Eun Soo Im: “Method for Tracking Multiple Objects by Using Multiple Cameras Based on Temporal-Spatial Graph Neural Network Utilizing Geometric Features for Each of the Multiple Objects and Object Tracking Device Using the Same.”

The invention describes a method in which an object-tracking device obtains descriptors for objects detected across camera views. It calculates similarities between descriptor pairs and uses a graph neural network to merge nodes that represent the same object, first forming a spatial graph and then a temporal-spatial graph. The resulting correspondence information supports tracking multiple objects within a defined space.

Career Highlights

Im’s listed patent reflects work on applying graph neural networks and geometric features to multi-camera object tracking. The approach combines spatial relationships among detected objects with temporal relationships across frames to help maintain object identities.

Collaborations

The supplied information identifies Jungkwon Lee as a coworker. It does not specify a particular collaboration between Lee and Im on the listed patent.

Conclusion

Eun Soo Im’s listed invention focuses on using temporal-spatial graph neural networks to connect object detections across cameras and support consistent multi-object tracking.