23 results for Sarrafzadeh, Hossein, Conference paper

  • Melanoma Diagnosis by the Use of Wavelet Analysis based on Morphological Operators

    Fassihi, Nima; Shanbehzadeh, Jamshid; Sarrafzadeh, Hossein; Ghasemi, Elham (2011-03)

    Conference paper
    Unitec

    Skin melanoma is the most dangerous type of skin cancer which is curable if diagnosed at the right time.. Drawing distinction between melanoma and mole is a difficult task and needs detailed laboratory tests. Utilizing morphologic operators in segmenting and wavelet analysis in order to extract the features has culminated in better result in melanoma diagnosis. This paper employs coefficients of wavelet decomposition to extract image's features. Melanoma classification is carried out by using the variance and mean of wavelet coefficients of images as the inputs of neural network. Results show 90% ability In distinction between benign and malignant lesions.

    View record details
  • An Alternative Approach for Developing Socially Assistive Robots

    Jayawardena, Chandimal; Kuo, I-Han; Sarrafzadeh, Hossein (2014)

    Conference paper
    Unitec

    This paper presents the design of the socially assistive companion robotic wheelchair named RoboChair. Unlike in most current companion robotics projects, the approach of RoboChair is not to build a completely new robotic device. Instead, the focus of the RoboChair project is to convert an already useful device (i.e. wheelchair) to a socially assistive companion robot. The authors argue that there are number of advantages in this approach. The proposed robotic chair is a mobile robot that can carry a person. It is equipped with several measuring devices for measuring vital signs. The robot chair is capable of engaging users with interactive dialogs through a touch screen and by using human-robot interaction techniques. It has a scalable modular software architecture so that adding new hardware and software modules is straightforward. The software frame- work is based on Robot Operating System (ROS) open source robotic middleware.

    View record details
  • Ensemble Statistical and Heuristic Models for Unsupervised Word Alignment

    Mohaghegh, Mahsa; Sarrafzadeh, Hossein; Mohammadi, Mehdi (2014)

    Conference paper
    Unitec

    Statistical word alignment models need large amounts of training data while they are weak in small-sized corpora. This paper proposes a new approach of an unsupervised hybrid word alignment technique using an ensemble learning method. This algorithm uses three base alignment models in several rounds to generate alignments. The ensemble algorithm uses a weighed scheme for resampling training data and a voting score to consider aggregated alignments. The underlying alignment algorithms used in this study include IBM Model 1, 2 and a heuristic method based on Dice measurement. Our experimental results show that by this approach, the alignment error rate could be improved by at least 15% for the base alignment models.

    View record details
  • Smart Task Orderings for Active Online Multitask Learning

    Pang, Paul; An, Jianbei; Zhao, Jing; Li, Xiaosong; Ban, Tao; Inoue, Daisuke; Sarrafzadeh, Hossein (2014-04-26)

    Conference paper
    Unitec

    This paper promotes active oMTL (i.e., Online Multitask Learning with task selection) by proposing two smart task ordering approaches: QR-decomposition Ordering and Minimal-loss Ordering, in which the optimal sequence of tasks for oMTL is computed as the training data/tasks are being presented. Our experimental results on four real-world datasets show that the proposed task orderings outperform all existing task ordering approaches to active oMTL.

    View record details
  • Interfaces that adapt like humans

    Alexander, Samuel T.V.; Sarrafzadeh, Hossein (2004)

    Conference paper
    Unitec

    Whenever people talk to each other, non-verbal behaviour plays a very important role in regulating their interaction. However, almost all human-computer interactions take place using a keyboard or mouse – computers are completely oblivious to the non-verbal behaviour of their users. This paper outlines the plan for an interface that aims to adapt like a human to the non-verbal behaviour of users. An Intelligent Tutoring System (ITS) for counting and addition is being implemented in conjunction with the New Zealand Numeracy Project. The system’s interface will detect the student’s non-verbal behaviour using in-house image processing software, enabling it to adapt to the student’s non-verbal behaviour in similar ways to a human tutor. We have conducted a video study of how human tutors interpret the non-verbal behaviour of students, which has laid the foundation for this research.

    View record details
  • The global cyber security workforce : an ongoing human capital crisis

    Fourie, Leon; Pang, Shaoning; Kingston, Tamsin; Hettema, Hinne; Watters, Paul; Sarrafzadeh, Hossein (2014-07)

    Conference paper
    Unitec

    Cyber threats pose substantial risk to government, businesses and individuals. There is an alarming shortage of trained professionals and academic programs to train and produce these professionals. Many countries including the US and New Zealand see this as a human capital crisis. This paper discusses the severity and various dimensions of this crisis. An analysis of data from a cyber security research centre collected on cyber attacks is presented, practical solutions are proposed and examples from New Zealand are detailed in this paper. The paper draws conclusions based on the comparative data collected on New Zealand and Japan.

    View record details
  • Eve : an affect-sensitive pedagogical agent

    Fourie, Leon; Overmyer, Scott; Shanbehzadeh, Jamshid; Sarrafzadeh, Hossein; Alexander, Samuel T.V. (2014-07)

    Conference paper
    Unitec

    Easy with Eve is an affect-sensitive tutoring system that features Eve, an animated pedagogical agent that adapts to student emotion based on the results of an observational study of human tutors. This paper presents the methodology and results of a study of the effectiveness of Eve, showing evidence that facial expression detection may have enhanced the effect of the presence of Eve on student motivation and beliefs about how much they were able to learn by interacting with the tutoring system, and suggesting areas of improvement to enhance the effectiveness of this system in business training and other context.

    View record details
  • Improved learning through peer tutoring in a declarative programming course

    Alexander, Samuel T.V.; Sarrafzadeh, Hossein (2014-07)

    Conference paper
    Unitec

    A method of peer tutoring was used and evaluated in an introductory programming course in Haskell. Students were paired on the basis of skill level with pairings opposite those suggested by the existing literature. This method of peer tutoring was shown to be effective in increasing the performance of both high- and low-skilled students in learning a declarative language. This method of peer tutoring can be applied to business and other technology education.

    View record details
  • Brightness preserving fuzzy dynamic histogram equalization

    Sarrafzadeh, Hossein; Rezazadeh, Fatemeh; Shanbehzadeh, Jamshid (2013)

    Conference paper
    Unitec

    Abstract—Image enhancement is a fundamental step of image processing and machine vision to improve the quality of an image for a specific application. Histogram equalization is an attractive and commonly-employed image enhancement algorithm which is used in certain circumstances because of its global nature. Brightness Preserving Dynamic Histogram Equalization (BPDHE) overcomes this problem by considering the local image histogram. However, this algorithm can result in false countering and ignoring of details. False countering is the result of dedicating wide intervals to intensities with high probability; ignoring details results from the wide distribution of regions with detailed information in small regions. This paper introduces a fuzzy version of BPDHE (i.e., BPFDHE) to overcome the aforementioned problems. The fuzzification is employed to provide a crisper version of an interval and of the number of pixels in that interval. This algorithm has been tested on 30 images under several different conditions. The results with BPFDHE, in terms of subjective quality, outperform histogram equalization and BPDHE.

    View record details
  • Fuzzy C-means based on automated variable feature weighting

    Nazari, Mousa; Shanbehzadeh, Jamshid; Sarrafzadeh, Hossein (2013)

    Conference paper
    Unitec

    Fuzzy C-means (FCM) is a powerful clustering algorithm and has been introduced to overcome the crisp definition of similarity and clusters. FCM ignores the importance of features in the clustering process. This affects its authenticity and accuracy. We can overcome this problem by appropriately assigning weights to features according to their clustering importance. This paper, proposes an improved FCM algorithm based on the method proposed by Huang by automated feature weighting. The simulation results on several UCI databases show that the proposed algorithm exhibits better performance than FCM.

    View record details
  • Designing a socially assistive companion robotic wheel chair: RoboChair

    Jayawardena, Chandimal; Baghaei, Nilufar; Ganeshan, Kathiravelu; Sarrafzadeh, Hossein (2013)

    Conference paper
    Unitec

    Developing socially assistive robots is an emerging interdisciplinary research area, which requires collaboration between a wide range of disciplines. Among recent research projects, there have been attempts to develop assistive robotic solutions to solve various health and social issues. In most current research attempts to design socially assistive robots, the focus is on designing new robotic agents that can interact with people by various means. Since people do not have much experience with robots, usually extensive field trials are conducted in order to assess the usability of these robots. However, determining the usability of these robotic agents is a difficult task, since the results of field trials are not always conclusive. In this paper, an attempt to overcome the difficulty of evaluating usability of assistive robots is presented. This paper presents the design of the first version of a companion robot called RoboChair. RoboChair is in the form of a wheel chair. But, functionally it is a socially assistive companion robot. The proposed robotic chair is a mobile robot that can carry a person. It is equipped with several measuring devices for taking clinical measurements mentioned above. In addition to that, it is equipped with several sensors for obstacle avoidance, map building, localization, detecting humans etc. It is also equipped with motor controllers and other actuators for motion control. The robot chair is capable of engaging users with interactive dialogs through a touch screen and by using human-robot interaction techniques. It has a scalable modular architecture so that adding new hardware and software modules is straightforward. The software framework is based on Robot Operating System (ROS) open source robotic middleware. RoboChair is controlled by a distributed controller that spans multiple hardware devices and multiple operating systems.

    View record details
  • An overview of different binary methods for documents based on their features

    Fazeli, Fatemeh; Sarrafzadeh, Hossein; Shanbehzadeh, Jamshid (2013)

    Conference paper
    Unitec

    This paper surveys binarization of document images. The main role of binarization is dimension and noise reduction. Binarization is one of the most important steps in preprocessing of document image understanding and compression. Image binarization means to classify image pixels into two classes, background and foreground. The input of this classification is a feature vector based on intensity values of image pixels. The new features are extracted from the first input vector and, according to the extracted features a cost function as a classifier is constructed. The intensity value that maximizes the cost function is considered as the boundary line of two classes. This paper divides the binarization algorithms into three groups. The first considers one input feature vector including intensity values of each pixel. The second one considers an input feature vector for each pixel based on the intensity value of the pixel and its neighbors. The third group is based on a combination of the first and second group of schemes.

    View record details
  • An Investigation of Factors and Measurements for Successful e-Commerce Websites

    Li, Xiaosong; Lian, Wei; Sarrafzadeh, Hossein (2014)

    Conference paper
    Unitec

    Business-to-Consumer (B2C) e-commerce is popular because of its convenience, speed and price. Although there has been intense debate about quality dimensions of e-commerce websites, more research is needed to find a well-established measurement. This empirical study identifies a set of measurements with 10 factors and their corresponding dimensions, including software development attributes based on the literature and the qualitative and quantitative data gathered from four different stakeholders. The survey results suggest that security, smooth transaction processes and smooth shopping processes are the most important concerns for online shoppers. The IS success model checking suggests that the proposed measurements are comprehensive. This work is compared with the customized ISO 9126 quality model.

    View record details
  • Cross-layer Optimisation for Quality of Service Support in Wireless Sensor Networks

    Mohaghegh, Mahsa; Manford, Chris; Sarrafzadeh, Hossein (2011)

    Conference paper
    Unitec

    Wireless sensor networks need to deliver real-time services such as video, audio and traditional data services therefore providing efficient quality of services (QoS) support is essential. In this paper we aim to address the time-delay parameter of QoS this is implemented using a new cross-layer framework design. The concept of cross-layer design is based on architecture where different layers can exchange information in order to improve the overall network performance. Promising results achieved by cross-layer optimization initiated significant research activity in this area. We present results from simulations of the new cross layer design and traditional OSI model using the OMNET++ software. We show that the cross layer design provides a feasible and flexible approach to solving the conflict between different layers in a standard OSI model. We demonstrate that the cross layer optimization is a promising solution and that enhances the quality of service in wireless sensor network applications.

    View record details
  • Skew detection of scanned document images

    Rezaei, , Sepideh Barekat; Sarrafzadeh, Hossein; Shanbehzadeh, Jamshid (2013)

    Conference paper
    Unitec

    Skewing of the scanned image is an inevitable process and its detection is an important issue for document recognition systems. The skew of the scanned document image specifies the deviation of the text lines from the horizontal or vertical axis. This paper surveys methods to detect this skew in two steps, dimension reduction and skew estimation. These methods include projection profile analysis, Hough Transform, nearest neighbor clustering, cross-correlation, piece-wise painting algorithm, piece-wise covering by parallelogram, transition counts, morphology.

    View record details
  • Ensemble learning methods for decision making : status and future prospects

    Ali, Shahid; Tirumala, Sreenivas Sremath; Sarrafzadeh, Hossein (2015-07)

    Conference paper
    Unitec

    In real world situations every model has some weaknesses and will make errors on training data. Given the fact that each model has certain limitations, the aim of ensemble learning is to supervise their strengths and weaknesses, leading to best possible decision in general. Ensemble based machine learning is a solution of minimizing risk in decision making. Bagging, boosting, stacked generalization and mixture of expert methods are the most popular techniques to construct ensemble systems. For the purpose of combining outputs of class labels, weighted majority voting, behaviour knowledge space and border count methods are used to construct independent classifiers and to achieve diversity among the classifiers which is important in ensemble learning. It was found that an ideal ensemble method should work on the principle of achieving six paramount characteristics of ensemble learning; accuracy, scalability, computational cost, usability, compactness and speed of classification. In addition, the ideal ensemble method would be able to handle large huge image size and long term historical data particularly of spatial and temporal. In this paper we reveal that ensemble models have obtained high acceptability in terms of accuracy than single models. Further, we present an analogy of various ensemble techniques, their applicability, measuring the solution diversity, challenges and proposed methods to overcome these challenges without diverting from the original concepts.

    View record details
  • Adaptive background modeling for land and water composition scenes

    Zhao, Jing; Pang, Shaoning; Hartill, Bruce; Sarrafzadeh, Hossein (2015-09)

    Conference paper
    Unitec

    In the context of maritime boat ramps surveillance, this paper proposes an Adaptive Background Modeling method for Land and Water composition scenes (ABM-lw) to interpret the traffic of boats passing across boat ramps. We compute an adaptive learning rate to account for changes on land and water composition scenes, in which the portion of water changes over time due to tidal dynamics and other environmental influences. Experimental comparative tests and quantitative performance evaluations of real-world boat-flow monitoring traffic sequences demonstrate the benefits of the proposed algorithm.

    View record details
  • A Hierarchical Phrase-Based Model for English-Persian Statistical Machine Translation

    Mohaghegh, Mahsa; Sarrafzadeh, Hossein (2012)

    Conference paper
    Unitec

    In this paper we show that a hierarchical phrasebased translation system will outperform a classical (nonhierarchical) phrase-based system in the English-to-Persian translation direction, yet for the Persian-to-English direction, the classical phrase-based system is preferable. We seek to explain why this is so, and detail a series of translation experiments with our SMT system using various bilingual corpora each with both toolkits Moses (non-hierarchical) and Joshua (hierarchical).

    View record details
  • A Text Localization Algorithm in Color Image via New Projection Profile

    Sarrafzadeh, Hossein; Aghajari, G; Shanbehzadeh, Jamshid (2010)

    Conference paper
    Unitec

    Text data in image present useful information for automatic annotation, indexing and structuring of images. In this paper, we propose an approach to automatically localize horizontally texts appearing in color and complex images. First, an edge detection method using a wavelet transform is used to finding text in image. Second, the image is binarized. Third, a new filter is applying for removing disperses pixels and non text area. After that, a new projection profile is applying for estimating text regions. The experimental results show that the proposed method achieves a much higher accuracy than existing methods. The advantage of this algorithm is low computation for finding text.

    View record details
  • Bionic voice (pilot study) : natural speech restoration for voice impaired individuals

    Sharifzadeh, Hamid; Allen, Jacqui; Sarrafzadeh, Hossein; Ardekani, Iman (2016-11)

    Conference paper
    Unitec

    The human voice is the most magnificent instrument for communication, capable of expressing deep emotions, conveying oral history through generations, or of starting a war. However, those who suffer from aphonia (no voice) and dysphonia (voice disorders) are unable to make use of this critical form of communication. They are typically unable to project anything more than hoarse whispers. Epidemiologic studies of the prevalence of voice disorders in the general adult population are rare. Nevertheless, information from a number of studies suggests that one third of the population have suffered from a temporary vocal impairment at some point in their life and that voice disorders can affect any age group and either sex. In some cases, vocal change is temporary however in those treated for malignant disease or with severe trauma there may be long term disturbance of phonation. This may affect occupation, social function and quality of life. Within a speech processing framework, we have worked on a novel method to return natural voice to laryngectomised people. This method leverages on recent advances in speech synthesis to deliver aworld-first technology. As a pilot study, this project has assessed the acoustic features of laryngectomised speech and has developed required enhancement for natural speech regeneration.

    View record details