Haifa, Israel

Alexander Bronstein

This inventor holds 29 USPTO granted patents and 4 published patent applications and 2 EPO patents, primarily in Image Processing Techniques. Top assignees: Technion Research & Development Foundation Ltd, Intel Corporation, Tel Aviv University. Active years: 2005-2024.

USPTO Granted Patents = 29 

% Patents Active = 65.5

 

Average Co-Inventor Count = 3.3

ph-index = 9

Forward Citations = 184(Granted Patents)


Location History:

  • Karmiel, IL (2013)
  • San Jose, CA (US) (2012 - 2014)
  • Haifa, IL (2005 - 2024)

Company Filing History:


Years Active: 2005-2024

Loading Chart...
Loading Chart...
Areas of Expertise:
Neural Network
Image Processing
Ultrasonic Imaging
Video Content Management
Facial Recognition
Distance Maps
Video Digest Generation
Spatio-Temporal Imaging
Multidimensional Scaling
Interactive Video
Coded Light Range Camera
Surface Similarity
29 patents (USPTO):Explore Patents

Title: Alexander Bronstein: A Pioneer in Neural Network Innovations

Introduction

Alexander Bronstein, based in Haifa, Israel, is a prominent inventor with a remarkable portfolio of 29 patents. His work focuses primarily on innovations in neural networks and image processing, showcasing his commitment to advancing technology in these fields.

Latest Patents

Bronstein’s latest patent, titled "System and method for emulating quantization noise for a neural network," introduces a sophisticated approach to training quantized neural networks. This system utilizes at least one hardware processor that receives input data consisting of multiple training input value sets and target value sets. In each training iteration, the system computes transformed values for each layer of the neural network by applying emulated non-uniformly quantized transformations, which include adding uniformly distributed random noise. This innovative approach effectively updates weight values to minimize the loss function, significantly enhancing neural network training efficiency.

Another significant patent from Bronstein is the "Method and system for processing an image." This technique dissects an image into various channels, each characterized by distinct depths-of-field. It leverages computer-readable mediums that store in-focus and out-of-focus dictionaries, computing sparse representations of the decomposed images. This method represents a breakthrough in image processing technology, advancing the ability to manipulate visual data effectively.

Career Highlights

Throughout his career, Alexander Bronstein has contributed significantly to research and development, notably working with the Technion Research & Development Foundation Limited. His inventive efforts have led to numerous patents that push the boundaries of current technology, particularly in the realms of neural networks and image processing.

Collaborations

Bronstein has collaborated with renowned colleagues Michael Bronstein and Ron Kimmel, furthering his work through comprehensive exchanges of innovative ideas and methodologies. This teamwork has enriched their research, contributing to the enhanced effectiveness of their projects and extending the impact of their patents.

Conclusion

Alexander Bronstein's contributions to technology, particularly in neural networks and image processing, have solidified his reputation as an influential inventor. With a robust collection of patents and a proven track record of innovative collaborations, he continues to shape the future of technology with his pioneering ideas and solutions.

Profile summary based on public USPTO records.
Please report any incorrect information to [email protected]
Loading…