Hamid Shojanazeri - Meta | LinkedIn (original) (raw)

About

I lead the AI Frameworks Partner Engineering team at Meta, focused on the ecosystem of…

Activity

Experience & Education

View Hamid’s full experience

See their title, tenure and more.

Licenses & Certifications

Publications

Image and Vision Computing New Zealand (IVCNZ) November 20, 2018

Similarity measure is an important research topic in image classification and retrieval. Given a type of image features, a good similarity measure should be able to retrieve similar images from the database while discarding irrelevant images from the retrieval. Similarity measures in literature are typically distance based which measure the spatial distance between two feature vectors in high dimensional feature space. However, this type of similarity measures do not have any perceptual meaning…
Similarity measure is an important research topic in image classification and retrieval. Given a type of image features, a good similarity measure should be able to retrieve similar images from the database while discarding irrelevant images from the retrieval. Similarity measures in literature are typically distance based which measure the spatial distance between two feature vectors in high dimensional feature space. However, this type of similarity measures do not have any perceptual meaning and ignore the neighborhood influence in the similarity decision-making process. In this paper, we propose a novel dissimilarity measure, which can measure both the distance and perceptual similarity of two image features in feature space. Results show the proposed similarity measure has a significant improvement over the traditional distance-based similarity measure commonly used in literature.
Other authors

Image and Vision Computing New Zealand (IVCNZ) November 20, 2018

Clustering similar images is an important task in image processing and computer vision. It requires a measure to quantify pairwise similarities of images. The performance of the clustering algorithm depends on the choice of the similarity measure. In this paper, we investigate the effectiveness of data independent (distance-based), data-dependent (mass-based) and hybrid (dis)similarity measures in the image clustering task using three benchmark image collections with different sets of features.…
Clustering similar images is an important task in image processing and computer vision. It requires a measure to quantify pairwise similarities of images. The performance of the clustering algorithm depends on the choice of the similarity measure. In this paper, we investigate the effectiveness of data independent (distance-based), data-dependent (mass-based) and hybrid (dis)similarity measures in the image clustering task using three benchmark image collections with different sets of features. Our results of K-Medoids clustering show that uses the hybrid Perceptual Dissimilarity Measure (PMD) produces better clustering results than distance-based Lp-based and mass-based mp-dissimilarity.
Other authors

Digital Image Computing: Techniques and Applications (DICTA) December 1, 2017

In image retrieval, an effective dissimilarity (or similarity) measure is required to retrieve the perceptually similar images. Minkowski-type distance is widely used for image retrieval, however it has its limitation. It focuses on distance between image features and ignores the data distribution of the image features, which can play an important role in measuring perceptual similarity of images. To address this limitation, a data dependent measure named m_p, which calculates the dissimilarity…
In image retrieval, an effective dissimilarity (or similarity) measure is required to retrieve the perceptually similar images. Minkowski-type distance is widely used for image retrieval, however it has its limitation. It focuses on distance between image features and ignores the data distribution of the image features, which can play an important role in measuring perceptual similarity of images. To address this limitation, a data dependent measure named m_p, which calculates the dissimilarity using the data distribution rather than geometric distance has been proposed recently. It considers two instances in a sparse region to be more similar than in a dense region. Relying only on data distribution and completely ignoring the geometric distance raise other limitations. This may result in finding two perceptually dissimilar instances similar due to being located in a sparse region or vice versa. We proposed a new hybrid dissimilarity measure and experimental results show that it addresses these limitations.
Other authors
See publication

Springer, Multimedia Tools and Applications

Optical and Quantum Electronics.

Courses

-

-

-

-

Projects

Jun 2018

Developed a clustering framework to use PDM in k-medoids. The clustering results of 11 benchmark datasets show the considerable improvement of using PDM as the dissimilarity measure for clustering compared with Euclidean distance and mass-based dissimilarity in k-medoids. Also k-medoids using PDM could outperform the k-means which is widely used in many scientific and industrial applications.

Sep 2017

A dissimilarity measure, which is more aligned with human perception, has been designed and developed. The new dissimilarity measure considers the effect of data distribution on the perceived dissimilarity by human. Different sets of features including HSV color histograms, LBP and SIFT BOW has been used to represent eBay, Texture and Corel datasets. The proposed dissimilarity measure outperformed Euclidean distance and mass-based dissimilarity measures.

Dec 2016

Considering the key role of an effective dissimilarity measure in accuracy of image retrievals, developed a new dissimilarity measure using the considering the geometric distance and data mass in measuring dissimilarity between image. Each of these components measures the dissimilarity from a different aspect. Corel and Caltech101 have been used as benchmark datasets and results show considerable improvement in retrievals.

Sep 2015

Developed an image retrieval and classification using Zernike moments features to represent images. Support Vector Machines (SVM) has been used to perform the classification task. Unlike the previous works that cited Zernike moments as a good candidate for image representation, our results revealed the limitations of these features.

2014

In recent years, there have been significant improvements in video technologies, which increase popularity of video applications. In our daily life, we face many advanced video applications by means of the recent technological developments. Those advanced video applications, which include interactive television, video conferencing, video search engines on the Internet, education, surveillance, sports, medical and news videos, require storing, archiving and querying of video files. People are…
In recent years, there have been significant improvements in video technologies, which increase popularity of video applications. In our daily life, we face many advanced video applications by means of the recent technological developments. Those advanced video applications, which include interactive television, video conferencing, video search engines on the Internet, education, surveillance, sports, medical and news videos, require storing, archiving and querying of video files. People are usually interested in the rich semantic information hidden in video files.Conventional database models are not adequate for handling video database applications since semantic video modeling and querying require additional techniques. This project concerns about a high semantic video retrieval to enhance the video search engines.

2013

Internet development and digital image modification software invited illegal access and usage of digital images, after which, digital watermarking emerged as a unique tool to protect the authenticity of digital images. This technique involves the insertion of an imperceptible message within the media. This project addressed problem with image authentication due to malicious forgery.
Other creators

2012

The advancement of Internet services and various storage technologies made video piracy as an increasing problem particularly with the proliferation of media sharing through the internet . Thus, research in copyright protection mechanisms and content authentication, where one of which includes digital watermarking has been receiving an increasing interest from scientists especially in designing a seamless algorithm for effective implementation. Basically digital watermarking…
The advancement of Internet services and various storage technologies made video piracy as an increasing problem particularly with the proliferation of media sharing through the internet . Thus, research in copyright protection mechanisms and content authentication, where one of which includes digital watermarking has been receiving an increasing interest from scientists especially in designing a seamless algorithm for effective implementation. Basically digital watermarking involves
embedding secret symbols known as watermarks within video data which can be used later for copyright detection and authentication verification purposes. This project address the concerns with content authentication and copyright protection of multimedia an intellectual properties.
Other creators

2010

Video piracy has become an increasing problem particularly with the proliferation of media sharing through the advancement of internet services and various storage technologies.
Thus, research in copyright protection mechanisms, where one of which includes digital watermarking has been receiving an increasing interest from scientists especially in designing a
seamless algorithm for effective implementation. Basically digital watermarking involves embedding secret symbols known as…
Video piracy has become an increasing problem particularly with the proliferation of media sharing through the advancement of internet services and various storage technologies.
Thus, research in copyright protection mechanisms, where one of which includes digital watermarking has been receiving an increasing interest from scientists especially in designing a
seamless algorithm for effective implementation. Basically digital watermarking involves embedding secret symbols known as watermarks within video data which can be used later for copyright
detection purposes. This project provides a critical review on various available techniques and addresses the main key performance indicators which include robustness, speed, capacity, fidelity, imperceptibility and computational complexity.
Other creators

2008

Today accounting software are essential need of any company, store department and any commercial enterprise.This project carried out the task of system analysis, database design and programming using Microsoft SQL Server and C# .NET to perform expected tasks of an accounting software.

Aug 2018 - Sep 2018

• Created a pipeline in python for real-time detection of objects.
• Created Scripts to perform video processing using OpenCV.
• Perform object detection using MobileNet and SSDs.

Jul 2018 - Aug 2018

• Designed and implemented a parking spot detection using YOLOV3.
• Created scripts for video processing and localization of a specific parking spot.
• Performed object detection using Yolov3.
• Created scripts to calculate the similarity of ROIs in different frames.

Jan 2018 - Mar 2018

• Created an OOP architecture to enable the use of different layers, loss functions, batch norm, dropout, and gradient descent algorithms.
• Wrote vectorized implementations for forward and backpropagation.
• Classification of the CIFAR-10 dataset using multi-layer Fully Connected Networks and CNNs.
• Case Studies of existing CNN models: AlexNet, ZFNet, and VGGNet.

Jan 2016 - May 2016

• Created scripts to detect color objects in videos using OpenCV.
• Created scripts for video processing using OpenCV.
• Created scripts to draw the object trajectory present in the video.

Honors & Awards

University Putra Malaysia

Feb 2013

University Putra Malaysia

Feb 2012

University Putra Malasyia

Sep 2011

University Putra Malaysia

Sep 2010

Organizations

Member

2013 - Present

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