Cheikhou BA

Title: Cheikhou BA and Image-Based Sign Detection

Introduction

Cheikhou BA is an inventor associated with Bentley Systems, Inc. His patent work addresses how software can detect and identify signs in images, combining machine-learning methods with image and text analysis.

Latest Patents

The provided information lists 1 patent for Cheikhou BA: “Techniques for Detecting and Identifying Signs in Images.” The invention describes a software application that detects physical objects such as signs and assigns them to broader meta-classes, while separating that detection process from identifying each object’s specific technical name.

The approach may use weakly supervised training with images labeled by meta-class to train a machine-learning detection model. After a sign is assigned a meta-class, the system can combine textual clues from optical character recognition with visual clues from keypoint matching of image descriptors to identify it.

Career Highlights

BA’s listed patent connects his work with image understanding, object detection, and machine learning. The invention’s staged approach distinguishes locating and categorizing signs from determining their specific identities.

Collaborations

The supplied information associates BA with Bentley Systems, Inc., and names Israël Akobi as a coworker. It does not specify their roles in developing the patent or describe a direct collaboration between them.

Conclusion

Cheikhou BA’s listed invention presents a method for detecting and identifying signs in images by combining weakly supervised machine learning, text recognition, and visual matching.