Frank Edughom Ekpar

Title: Frank Edughom Ekpar: Inventor of a Neural-Network-Based Panoramic Image Navigation System.

Introduction

Frank Edughom Ekpar is an inventor based in Toyama, Japan. His work includes a patented system that uses a neural network to correct distortion in wide-angle images and support navigation through panoramic views.

Latest Patents

Ekpar is associated with 1 patent: “Panoramic Image Navigation System Using Neural Network for Correction of Image Distortion.” The invention enables users to navigate a wide-angle image by generating undistorted sub-images.

The system is trained using a calibration pattern placed within the field of view of a wide-angle imaging apparatus. It uses the calibration points’ view directions and their positions in the captured image to teach a neural network how to map viewing directions to image locations. When generating a sub-image, the system converts display-pixel positions into view directions, maps those directions to positions in the original image, and assigns the corresponding image attributes to the sub-image’s pixels.

Career Highlights

Ekpar’s listed patent reflects work at the intersection of neural networks, image processing, and panoramic navigation. The invention addresses distortion correction by learning the relationship between viewing directions and positions within a wide-angle image.

Collaborations

Ekpar is associated with Tateyama R & D Co., Ltd. His listed coworker is Shinya Yamauchi.

Conclusion

Frank Edughom Ekpar’s patented system applies neural-network-based mapping to make navigation through distorted wide-angle and panoramic images more practical.