Johannes Peter Kopf

Title: Innovations of Johannes Peter Kopf

Introduction

Johannes Peter Kopf is an accomplished inventor based in Medina, WA (US). He holds a total of four patents, showcasing his expertise in the field of technology and innovation. His work primarily focuses on advancements in neural networks and video depth estimation.

Latest Patents

One of his latest patents is titled "Learning neural light fields with ray-space embedding networks." This invention involves a computing system that accesses a set of training images to train a neural light field network for view synthesis. The system casts rays into a three-dimensional space for each pixel of the training image, transforming ray coordinates and generating predicted color values for pixels. This process includes comparing predicted values with ground-truth values to update the network.

Another significant patent is "Robust consistent video depth estimation." In this invention, a computing system receives a video comprising multiple image frames. It generates initial depth maps using a machine-learning model and computes misalignment errors. The system then optimizes camera poses and generates refined depth maps by adjusting initial maps based on flexible deformation splines.

Career Highlights

Johannes has worked with notable companies such as Meta Platforms Technologies, LLC and Meta Platforms, Inc. His contributions to these organizations have significantly advanced the fields of machine learning and computer vision.

Collaborations

Throughout his career, Johannes has collaborated with talented individuals, including Jia-Bin Huang and Changil Kim. These partnerships have fostered innovation and creativity in his projects.

Conclusion

Johannes Peter Kopf is a prominent inventor whose work in neural networks and video depth estimation has made a significant impact in technology. His patents reflect his dedication to advancing the field and improving computational methods.