Tara Akhavan, Ph.D.

EY Entrepeneur Of The Year 2019 Finalist

Dr. Akhavan is a visionary, entrepreneur and pioneer in her field, Computer Vision Machine Learning (CVML). From inception, she led the development of Perceptual Display Platform (IRYStec PDP) software, a game-changing technology in dynamic content optimization that brings displays and content into alignment with human vision. The adoption of this new technology will transform our display viewing experience making displays easier to read, healthier, safer and more power efficient. Dr. Tara Akhavan is the Chairwoman company technology committee that synchronizes research (applied/basic) with 12+ university professors.

Prior to founding IRYStec, Dr. Akhavan was awarded for scaling an Operations and Maintenance Center (OMC) product in the Telecommunications industry, all the way from analysis and design to deployment in a 3GPP mobile network with 20 Million subscribers.

Dr. Akhavan is shining example of how Canada attracts not only the best talent in the world, but also the most diverse talent. She has only been in Canada since late 2014 and was invited to speak at TEDx Women in Montreal in 2016. She is an active promoter of female leaders in different technology forums and volunteer in technology events. Dr. Akhavan is an active member of The Society for Information Display and serves as Marketing Vice Chair. In 2018 she led a panel on Women in Technology at the largest display conference in the world called Display Week. Dr. Akhavan has been recently named EY Entrepreneur Of The Year 2019 award finalist in Quebec.

Dr. Akhavan obtained a bachelor’s degree in Computer Engineering, a Master degree in Artificial Intelligence and a Ph.D. in Image Processing and Computer Vision from Vienna University of Technology.

Interviews

Tara Akhavan at Women of Influence
Dr. Tara Akhavan was panelist at the Women of Influence series dedicated to discuss new emerging technologies.
Tara Akhavan Display Week 2018
IRYStec Software showcased the Perceptual Display Platform (PDP) embedded software solutions at SID Display Week 2018.
Tara Akhavan - Taking on the World at Display Week 2018
Tara Akhavan - Taking on the World: Women in Tech Forum at Display Week 2018
Tara Akhavan - Women Techmakers Montréal
When I realized I need to keep up my personal growth with the business growth by Tara Akhavan
ScaleupFest 2017: Stories from the trenches (Chain Reaction Panel)
ScaleupFest 2017: Stories from the trenches (Chain Reaction Panel)
Tara_Akhavan Display Week 2017
Dr. Tara Akhavan participated in Display Week’s first “Women in Tech” forum. She shared her achievements in the display industry and talked about a new set of role models.

Publications

Exploiting Wide-Gamut Displays

By Gregory Ward, Hyunjin Yoo, Afsoon Soudi, Tara Akhavan
Society of Imaging Science and Technology (CIC 24)
Published: Nov 11, 2017

Abstract: We present a hybrid color mapping (HCM) designed to preserve a selected region in chromaticity space while exploiting the larger gamut of the intended target display. Our method is based on the hypothesis that outside a certain set of critical colors, people are less particular and prefer a more saturated appearance if available. This hypothesis was borne out in the subject study we conducted. While other practitioners have implemented similar gamut-mapping techniques, our definition of preserved colors is more flexible and not based purely on saturation level. We employ an exponent or “acceleration factor” in our mapping to better preserve neighboring colors for a more natural appearance. Our method further avoids contrast and luminance changes, works between arbitrary gamuts, and is a fully invertible one-to-one mapping between color volumes.

Irystec DriveSafe, Ambient Adaptive Software, Makes Driving Safer

By Afsoon Soudi, Mehdi Rezagholizadeh, Tara Akhavan
SID Symposium Digest of Technical Papers
Published: 25 May 2016

Abstract: DriveSafe is a real-time, ambient adaptive, perceptual software product. It creates a natural viewing experience taking into account the human visual system characteristics, computer vision research and the cutting-edge display technology. DriveSafe improves readability and makes reading vehicle displays safer in night time and bright light conditions.

A Retargeting Approach for Mesopic Vision: Simulation and Compensation

By Mehdi Rezagholizadeh, Tara Akhavan, Afsoon Soudi, Hannes Kaufmann, James J. Clark
Journal of Imaging Science and Technology
Society for Imaging Science and Technology
Published: Jan 7, 2016

Abstract: Retargeting approaches aim at providing a unified framework for image rendering in which both the intended scene luminance and the actual luminance of the display are taken into account. At the core of any color retargeting method, a color vision model and its inverse are employed. Such a color appearance model should be invertible and cover the entire luminance range of the human visual system. There are not many available models that meet these two conditions. Moreover, most of these models are developed based on psychophysical experiments over color patches, and many have never been used for complex images due to their complexity. In this article, a color retargeting approach based on the mesopic model of Shin et al. [“A color appearance model applicable in mesopic vision,” Opt. Rev. 11, 272–278 (2004)] is developed to work with complex images. The authors propose an inverse model for complex images to compensate for color appearance changes on dimmed displays viewed in a dark environment. Their experimental results using both quantitative and qualitative evaluations show a discriminative improvement in the perceived color quality for mesopic vision. The proposed method can be incorporated into image retargeting techniques and display rendering mechanisms.

Backward Compatible HDR Stereo Matching: A Hybrid Tone-Mapping Based Framework

By Tara Akhavan & Hannes Kaufmann
EURASIP Journal on Image and Video Processing
Published: Nov 14, 2015

Abstract: Stereo matching under complex circumstances, such as low-textured areas and high dynamic range (HDR) scenes, is an ill-posed problem. In this paper, we introduce a stereo matching approach for real-world HDR scenes which is backward compatible to conventional stereo matchers. For this purpose, (1) we compare and evaluate the tone-mapped disparity maps to find the most suitable tone-mapping approach for the stereo matching purpose. Thereof, (2) we introduce a combining graph-cut based framework for effectively fusing the tone-mapped disparity maps obtained from different tone-mapped input image pairs. And finally, (3) we generate reference ground truth disparity maps for our evaluation using the original HDR images and a customized stereo matching method for HDR inputs. Our experiments show that, combining the most effective features of tone-mapped disparity maps, an improved version of the disparity is achieved. Not only our results reduce the low dynamic range (LDR), conventional disparity errors by the factor of 3, but also outperform the other well-known tone-mapped disparities by providing the closest results to the original HDR disparity maps.

Evaluation of LDR, Tone Mapped and HDR Stereo Matching Using Cost-volume Filtering Approach

By Tara Akhavan, Hyunjin Yoo, Margrit Gelautz
European Signal Processing Conference (EUSIPCO 2014)
Published: Jun 1, 2014

Abstract: We present stereo matching solutions based on a fast cost volume filtering approach for High Dynamic Range (HDR) scenes. Multi-exposed stereo images are captured and used to generate HDR and Tone Mapped (TM) images of the left and right views. We perform stereo matching on conventional, Low Dynamic Range (LDR) images, original HDR, as well as TM images by customizing the matching algorithm for each of them. An evaluation on the disparity maps computed from the different approaches demonstrates that stereo matching on HDR images outperforms conventional LDR stereo matching and TM stereo matching, with the most discriminative disparity maps achieved by using HDR color information and log-luminance gradient values for matching cost calculation.

Read full text

Stereo HDR Disparity Map Computation Using Structured Light

By Tara Akhavan, Christian Kapeller, Ji-Ho Cho, Margrit Gelautz
HDRi2014 – Second International Conference and SME Workshop on HDR imaging
Published: Feb 21, 2014

Abstract: We present stereo matching solutions based on a fast cost volume filtering approach for High Dynamic Range (HDR) scenes. Multi-exposed stereo images are captured and used to generate HDR and Tone Mapped (TM) images of the left and right views. We perform stereo matching on conventional, Low Dynamic Range (LDR) images, original HDR, as well as TM images by customizing the matching algorithm for each of them. An evaluation on the disparity maps computed from the different approaches demonstrates that stereo matching on HDR images outperforms conventional LDR stereo matching and TM stereo matching, with the most discriminative disparity maps achieved by using HDR color information and log-luminance gradient values for matching cost calculation.

Read full text

A Framework for HDR Stereo Matching Using Multi-exposed Images

By Tara Akhavan, Hyunjin Yoo, Hyunjin Yoo, Margrit Gelautz
HDRi 2013: First International Conference and SME Workshop on HDR imaging
Published: Apr 10, 2013

Abstract: We present stereo matching solutions based on a fast cost volume filtering approach for High Dynamic Range (HDR) scenes. Multi-exposed stereo images are captured and used to generate HDR and Tone Mapped (TM) images of the left and right views. We perform stereo matching on conventional, Low Dynamic Range (LDR) images, original HDR, as well as TM images by customizing the matching algorithm for each of them. An evaluation on the disparity maps computed from the different approaches demonstrates that stereo matching on HDR images outperforms conventional LDR stereo matching and TM stereo matching, with the most discriminative disparity maps achieved by using HDR color information and log-luminance gradient values for matching cost calculation.

Read full text

A Color Constancy Method Using Fuzzy Measures and Integrals

By Tara Akhavan, Mohsen Ebrahimi Moghaddam
HDRi 2013: First International Conference and SME Workshop on HDR
Published: Mar 23, 2011

Abstract: The ability of measuring colors of objects, independent of light source illumination, is called color constancy which is an important problem in machine vision and image processing fields. In this paper, we propose a new combinational method that is based on fuzzy measures and integrals to estimate the chromaticity of the light source as the major step of color constancy. The basic idea of the proposed method is that there are color constancy methods with some similarities in their structure and the way they are applied. The proposed method works with the help of assigning fuzzy measures to these methods and their combinations and computing the Choquet fuzzy integral. To approve the proposed method, we selected four well known algorithms and their results were combined by the proposed approach. In selecting these methods, it was tried to choose the ones which had better performance in compare to other methods, however the proposed method can be applied on any other methods just by adjusting its parameters. It is shown in this article that proposed approach performs better than other proposed methods for color constancy most of the time.

Read full text

A new combining learning method for color constancy

By Tara Akhavan, Mohsen Ebrahimi Moghaddam
2nd International Conference on Image Processing Theory Tools and Applications (IPTA), 2010
Published: Jul 2010

Abstract: The ability to measure color of objects, independent of color of the light source, is called color constancy which is an important problem in machine vision and image processing. In this paper, we propose a method that employs a neural network to estimate the chromaticity of light source. This network uses the results of four well known color constancy methods as its input in training and tries to find the best result in test phase. In selecting the input methods, it has been tried to select ones which each one focuses on a particular specification of the colored image and is suitable for training also. By considering these issues, Max RGB, gray world assumption, gray edge, and shades of gray as well known methods were selected. In the proposed methods, the result in test phase may correspond with none of these algorithms necessarily. The experimental results showed that the proposed method reached to a good estimation of the illuminant source with less complexity in comparison to the previous related works.

Read full text

Developing a persian chunker using a hybrid approach

By Soheila Kiani, Tara Akhavan, Mehrnoush Shamsfard
IMCSIT’09-Computational Linguistics and Applications (CLA’09), Mragowa, Poland
Published: Oct 12, 2009

Abstract: Text segmentation is the process of recognizing boundaries of text constituents, such as sentences, phrases and words. This paper focuses on phrase segmentation also known as chunking. This task has different problems in various natural languages depending on linguistic features and prescribed form of writing. In this paper, we will discuss the problems and solutions especially for the Persian language and present our system for Persian phrase segmentation. Our system exploits a hybrid method for automatic chunking of Persian texts. The method at first exploits a rule-based approach to create a tagged corpus for training a neural network and then uses a multilayer perceptron neural network and Fuzzy C-Means Clustering to chunk new sentences. Experimental results show the average precision of %85.7 for the chunking result.

Read full text

Parsumist: A Persian text summarizer

By Mehrnoush Shamsfard, Tara Akhavan, Mona Erfani Jourabchi
IEEE International Conference on Natural Language Processing and Knowledge Engineering – NLP-KE , 2009
Published: Sep 2009

Abstract: The rapid growth of online information services causes the problem of information explosion. Automatic text summarization techniques are essential for dealing with this problem. The process of compacting a source document to reduce complexity and length, retaining the most important information is called text summarization. This paper introduces PARSUMIST; a text summarization system for Persian documents. It can generate generic or topic/query-driven extract summaries for single or multiple Persian documents, using a combination of statistical, semantic and heuristic improved methods. In this paper we will first review the related works in this field and especially in Persian text summarization. Then we will present the architecture of PARSUMIST, its components and its features. The last section will evaluate the system and compare it to other existing ones.

Read full text

Persian Document Summarization by Parsumist

By Mehrnoush Shamsfard, Tara Akhavan and Mona Erfani Joorabchi
World Applied Sciences Journal 7 (Special Issue of Computer & IT): 199-205, 2009
Published: 2009

Abstract: The rapid growth of online information services has created the problem of information
explosion. Automatic text summarization techniques are essential for dealing with this problem. The process of compacting a source document to reduce its complexity and length while retaining its most important contents is called text summarization. This paper introduces Parsumist-a text summarization system for Persian documents. It exploits a combination of statistical, semantic and heuristic-improved methods. It can generate generic or topic/query-driven extracts summaries for single-or multiple Persian documents. In this paper, we first review the related work in this field, especially for Persian text summarization. We then present the architecture of Parsumist, its components and features. The last section evaluates the system and compares it to other systems that exist.

Read full text