Bucharest, Romania

George Leotescu

This inventor holds 1 USPTO granted patent. Top assignee: Amazon Technologies, Inc.. Active years: 2026.

IDiyas Innovation Intelligence. (2026). Inventor Profile: George Leotescu. Retrieved from https://idiyas.com/inventor/george-leotescu

Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile


Average Co-Inventor Count = 1.0


Company Filing History:


Years Active: 2026

Loading Chart...
1 patent (USPTO):Explore Patents

Title: George Leotescu: Inventor of Machine-Learning Systems for Image Representations.

Introduction

George Leotescu is an inventor associated with Bucharest, Romania, and Amazon Technologies, Inc. His patent work includes a machine-learning approach for creating a unified representation from multiple images of an object.

Latest Patents

Leotescu is credited with 1 patent: “Machine Learning Systems for Generating Representations Based on Multiple Images of an Object.” The invention processes images sequentially, encoding each image and updating a shared latent representation through cross-attention. After all images have been processed, the resulting representation can emphasize characteristics common across the images while de-emphasizing features that are not shared. This representation may then support image-based search operations.

Career Highlights

The available information associates Leotescu’s patent activity with Amazon Technologies, Inc. His work described in the patent applies machine learning to combine visual information from multiple images into a representation useful for subsequent search.

Collaborations

Alin-Ionut Popa is identified as Leotescu’s coworker. The provided information does not specify whether Popa collaborated on this patent.

Conclusion

George Leotescu’s patent describes a method for consolidating information from multiple images into a single representation, supporting image-based search through machine learning.

Profile summary based on public USPTO records.
Data Sources: USPTO Patent Grant XML, Patent Center, EPO & CIPO • Normalized by IDiyas Innovation Graph. Methodology & provenance architecturePlease report any incorrect information to [email protected]
Loading…