Sean Albert Sedwards

Title: Sean Albert Sedwards and His Work in 3D Perception

Introduction

Sean Albert Sedwards is an inventor based in Kitchener, Canada, whose patent work focuses on three-dimensional (3D) point-cloud processing and object detection. His listed portfolio includes 2 patents, with inventions addressing training data generation and the construction of 3D scene representations.

Latest Patents

One of Sedwards’s latest patents, “Method and a Server for Generating Training Data for Training an Object Detector,” describes a way to create training data for fine-tuning an object detector. The method identifies a quasi-stationary bounding box in a sequence of labelled 3D point clouds, relocates an object’s bounding box to a corresponding position, and stores the relabelled point cloud. This approach is designed to support the preparation of training examples across different domains.

His other listed patent, “Methods and Systems for Generating a Three-Dimensional Representation of a Scene,” concerns combining information from successive 3D point clouds. It detects an object, estimates its position in a later point cloud, and aggregates data points representing that object from preceding point clouds. The resulting representation may help improve the accuracy of driving-scenario analysis.

Career Highlights

Sedwards’s listed inventions address practical challenges in computer vision and 3D perception. One focuses on generating data for object-detector training, while the other uses temporal information from point-cloud sequences to build scene representations.

Collaborations

Sedwards is associated with Huawei Technologies Canada Co., Ltd. His listed coworkers include Chengjie Huang and Vahdat Abdelzad.

Conclusion

Sean Albert Sedwards’s patent work centers on methods for preparing object-detection training data and generating 3D representations from point-cloud sequences, contributing to technologies relevant to scene understanding and driving analysis.