Lior Elazary

Lior Elazary

Agoura Hills, California, United States
2K followers 500+ connections

Join to view profile

About

...because this life is yours. Some of it was given to you, the rest you build yourself.

Activity

Join now to see all activity

Experience

Education

Licenses & Certifications

  • Ham Radio Operator

    FCC

    Issued
    Credential ID KK6BWA

Volunteer Experience

  • CS advisor

    Lindero Canyon Middle School

    - Present 10 years 11 months

  • Founder / Creating curriculums and teaching kids CS & Robotics

    http://techblazer.org/

    - Present 11 years 3 months

    Education

    TechBlazer is a non-profit organization with the goal of inspiring our kids to become technologists or tech entrepreneurs. Based on the YMCA Indian Guides model, it will involve collaboration between parents and their sons or daughters. Parents do not have to have a tech background, just an interest in technology. Depending on sign-up numbers, the group will be segmented in tribes of 10 or less. At least once a month, the group will meet to undertake a project.

Publications

  • A Bayesian model for efficient visual search and recognition

    Vision Research, Vol. 50, No. 14, pp. 1338-1352 (Made the most cited Vision Research Article list)

    Humans employ interacting bottom-up and top-down processes to significantly speed up search and recognition of particular targets. We describe a new model of attention guidance for efficient and scalable first-stage search and recognition with many objects (117,174 images of 1147 objects were tested, and 40 satellite images). Performance for recognition is on par or better than SIFT and HMAX, while being, respectively, 1500 and 279 times faster. The model is also used for top-down guided…

    Humans employ interacting bottom-up and top-down processes to significantly speed up search and recognition of particular targets. We describe a new model of attention guidance for efficient and scalable first-stage search and recognition with many objects (117,174 images of 1147 objects were tested, and 40 satellite images). Performance for recognition is on par or better than SIFT and HMAX, while being, respectively, 1500 and 279 times faster. The model is also used for top-down guided search, finding a desired object in a 5x5 search array within four attempts, and improving performance for finding houses in satellite images.

    Other authors
    • L. Itti
    See publication

  • Framework and implementation for perception

    In: Proc. Vision Science Society Annual Meeting

    A biologically-inspired framework for perception is proposed and implemented, which helps guide the systematic development of machine vision algorithms and methods. The core is a hierarchical Bayesian inference system. Hypotheses about objects in a visual scene are generated 'bottom-up'​ from sensor data. These hypotheses are refined and validated 'top-down' when complex objects, hypothesized at higher levels, impose new feature and location priors on the component parts of these objects at…

    A biologically-inspired framework for perception is proposed and implemented, which helps guide the systematic development of machine vision algorithms and methods. The core is a hierarchical Bayesian inference system. Hypotheses about objects in a visual scene are generated 'bottom-up'​ from sensor data. These hypotheses are refined and validated 'top-down' when complex objects, hypothesized at higher levels, impose new feature and location priors on the component parts of these objects at lower levels. To efficiently implement the framework, an important new contribution is to systematically utilize the concept of bottom-up saliency maps to narrow down the space of hypotheses. In addition, we let the system hallucinate top-down (manufacture its own data) at low levels given high-level hypotheses, to overcome missing data, ambiguities and noise. The implemented system is tested against images of real scenes containing simple 2D objects against various backgrounds. The system correctly recognizes the objects in 98.71\​% of 621 video frames, as compared to SIFT which achieves 38.00\​%.

    Other authors
    • L. Itti

  • Interesting objects are visually salient

    Journal of Vision, Vol. 8, No. 3:3, pp. 1-15

    How do we decide which objects in a visual scene are more interesting? While intuition may point toward high-level object recognition and cognitive processes, here we investigate the contributions of a much simpler process, low-level visual saliency. We used the LabelMe database (24,863 photographs with 74,454 manually outlined objects) to evaluate how often interesting objects were among the few most salient locations predicted by a computational model of bottom-up attention. In 43\​% of all…

    How do we decide which objects in a visual scene are more interesting? While intuition may point toward high-level object recognition and cognitive processes, here we investigate the contributions of a much simpler process, low-level visual saliency. We used the LabelMe database (24,863 photographs with 74,454 manually outlined objects) to evaluate how often interesting objects were among the few most salient locations predicted by a computational model of bottom-up attention. In 43\​% of all images the model's predicted most salient location falls within a labeled region (chance 21\​%). Furthermore, in 76\​% of the images (chance 43\​%), one or more of the top three salient locations fell on an outlined object, with performance levelling off after six predicted locations. The bottom-up attention model has neither notion of object nor notion of semantic relevance. Hence, our results indicate that selecting interesting objects in a scene is largely constrained by low-level visual properties rather than solely determined by higher cognitive processes.

Patents

  • Autonomous order fulfillment and inventory control robots

    Issued US 9120622

    Some embodiments provide a system for fully autonomous order fulfillment and inventory management within a distribution site or warehouse. The system operates by autonomous robots directly pulling customer order items from distribution site shelves or pulling individual bins from distribution site shelves and dispensing appropriate quantities of items from the bins until all items of a customer order are retrieved without human involvement. The system further involves the robots autonomously…

    Some embodiments provide a system for fully autonomous order fulfillment and inventory management within a distribution site or warehouse. The system operates by autonomous robots directly pulling customer order items from distribution site shelves or pulling individual bins from distribution site shelves and dispensing appropriate quantities of items from the bins until all items of a customer order are retrieved without human involvement. The system further involves the robots autonomously monitoring item quantities within the distribution site, identifying and autonomously responding to shortages, and organizing the items within the distribution site for most efficient order fulfilment.

    Other inventors

  • Human and robotic distributed operating system (HaRD-OS)

    Issued US 9050723

    Some embodiments provide a human and robot distributed operating system (HaRD-OS). The HaRD-OS efficiently and dynamically connects different human operators and algorithms to multiple remotely deployed robots based on the task(s) that the robots are to complete. Some embodiments facilitate an action/perception loop between the operators, algorithms and robots by routing tasks between human operators or algorithms based on various routing polices including operator familiarity, aptitude, access…

    Some embodiments provide a human and robot distributed operating system (HaRD-OS). The HaRD-OS efficiently and dynamically connects different human operators and algorithms to multiple remotely deployed robots based on the task(s) that the robots are to complete. Some embodiments facilitate an action/perception loop between the operators, algorithms and robots by routing tasks between human operators or algorithms based on various routing polices including operator familiarity, aptitude, access rights, end user preference, time zones, costs, efficiency, latency, privacy concerns, etc. In the event of a fault, some embodiments route to the best operator or algorithm that is able to handle the particular fault with the required privileges. Some embodiments secure task guarantees to be completed at particular times, priorities or costs.

    Other inventors
    See patent

  • Real-time granular statistical reporting for distributed platforms

    Issued US 8510807

    Some embodiments provide a reporting system for improved granular real-time performance statistics reporting in a distributed platform. The reporting system includes a statistic server and a portal. The statistics server is communicably coupled to servers of the distributed platform that produce statistical data related to the distribution of content and execution of services for different customers. The statistics server aggregates the statistical data from the plurality of servers in an…

    Some embodiments provide a reporting system for improved granular real-time performance statistics reporting in a distributed platform. The reporting system includes a statistic server and a portal. The statistics server is communicably coupled to servers of the distributed platform that produce statistical data related to the distribution of content and execution of services for different customers. The statistics server aggregates the statistical data from the plurality of servers in an optimized staggered manner during a recurring interval. This reduces the amount of statistical data that is passed at any particular instance in time from the servers the statistics servers. The statistics server incrementally updates a real-time performance report for a particular customer as the statistical data is aggregated for the particular customer so that the computational and memory overhead for deriving the performance report in real-time is reduce.

    Other inventors
    • Robert J. Peters
    • Seungyeob Choi

  • Content network global replacement policy

    Issued US 8095737

    This invention is related to content delivery systems and methods. In one aspect of the invention, a content provider controls a replacement process operating at an edge server. The edge server services content providers and has a data store for storing content associated with respective ones of the content providers. A content provider sets a replacement policy at the edge server that controls the movement of content associated with the content provider, into and out of the data store. In…

    This invention is related to content delivery systems and methods. In one aspect of the invention, a content provider controls a replacement process operating at an edge server. The edge server services content providers and has a data store for storing content associated with respective ones of the content providers. A content provider sets a replacement policy at the edge server that controls the movement of content associated with the content provider, into and out of the data store. In another aspect of the invention, a content delivery system includes a content server storing content files, an edge server having cache memory for storing content files, and a replacement policy module for managing content stored within the cache memory. The replacement policy module can store portions of the content files at the content server within the cache memory, as a function of a replacement policy set by a content owner.

    Other inventors
    • Alex Kazerani
    • Jay Sakata

Honors & Awards

  • National Computing Research Association's (CRA) 2005 Outstanding Undergraduate Awards competition.

    -

Languages

  • English

    -

  • Hebrew

    -

More activity by Lior

View Lior’s full profile

  • See who you know in common
  • Get introduced
  • Contact Lior directly
Join to view full profile

Other similar profiles

Explore more posts

Explore top content on LinkedIn

Find curated posts and insights for relevant topics all in one place.

View top content

Add new skills with these courses

See all courses