Menu:

Accepted Papers

  • The Comparison of SIFT and SURF Feature Detectors and Descriptors for Mobile Robot Visual Odometry
    Houssem Eddine Benseddik, Oualid Djekoune and Mahmoud Belhocine, Center for Development of Advanced Technologies, Algeria
    ABSTRACT
    This paper presents the evaluation of SIFT and SURF point detectors and descriptors in the context of monocular visual odometry for mobile robot. With this comparaison attempt to answer the question of choosing the most adapted features detectors and descriptors for estimating mobile robot egomotion. We evaluate the performance of these detectors/ descriptors from the repeatability, recall, precision and cost of computation. To estimate the relative pose of camera from outlier-contaminated feature correspondences, the essential matrix and inlier set is estimated using RANSAC. Experimental results demonstrate that SURF, outperform the SIFT detector/ descriptor, in terms of accuracy and speed.


  • Desktop Fireworks using Augmented Reality
    Aisha Urooj Khan and Komil Iqbal, Lahore University of Management & Sciences (LUMS), Pakistan
    ABSTRACT
    The ABSTRACT of our project is to make an application using Augmented Reality. Our focus was to learn and develop an application using augmented reality, which will give us a practical experience in this computer vision course. We learned basic computer vision concepts, which couldn’t be so better without making this project. Our main goal was to do video-in-video augmentation.


  • Automatic Flotation Froth Bubble Size Distribution Estimation Using Meanshift and Watershed Transform
    Anthony Amankwah1 and Chris Aldrich2, 1GIMPA, Ghana and 2Curtin University, Australia
    ABSTRACT
    Flotation is a process used in the mineral industry to concentrate valuable minerals. The efficiency of the flotation process is influenced by the size distribution of the flotation froth. In this work, we propose a new method for the automatic size distribution estimation using the mean shift and watershed algorithm. The mean shift algorithm is used to find pixel clusters of particular modes of the probability density function of the image data. The pixel clusters which are above a threshold as used as markers for the watershed transform. Experimental results show that the proposed method was fast and robust for determining the size distribution of different classes of flotation froth.


  • On Interval Estimating Regression
    Marcin Michalak, Silesian University of Technology, Poland
    ABSTRACT
    This paper presents a new look on the well-known nonparametrical regression estimator – the Nadaraya- Watson kernel estimator. After 50 years the application of uncertainty theory – interval analysis – is applied for estimator. The paper presents the background of Nadaraya-Watson kernel estimator and interval analysis basics and shows the interval Nadaraya-Watson kernel estimator.


  • Real Time Postural Control of Humanoid Robot based on Imitation Control
    Hee Jung Byun and Su Goog Shon, The University of Suwon, Korea
    ABSTRACT
    Imitation control is a well-known method for biped robot walking, which uses an operator motion data to a robot’s walk. But, direct usage of the data usually results in dynamically unstable motion. Also, time delay between the operator’s motion data and the robot’s joint motion through a control communication system makes the stability of the robot walking even worse. This paper studies challenging issues to overcome the postural instability caused by both control communication delays and dynamics of the robot. Key concept is to trace the movement of the center of pressure and not to fall down by control of robot joints as fast as possible. The experimental results demonstrate that the biped robot successfully imitates the operator motions and can walk stably through the real-time control for the robot’s ankle.


  • Research and application of spiral welded pipe information management system
    Jing Wang1,2, Jun Zhu1, Bo Shi1, Shengbo Rong1 and Huanyu Zhang1, 1Shenyang Institute of Automation Chinese Academy of Sciences and 2University of Chinese Academy of Sciences
    ABSTRACT
    Production of spiral welded pipe is a complex process that consists of multiple forming and inspecting stations, between different stations there often requires transmission and exchange of information. Traditional manufacturers for spiral welded pipe only have a low level of automation and informatizaion, production data, inspection data, devices status data and other information are manual recorded by paper reports, consuming lots of manpower and resources. Moreover for some stations, such as X-ray inspection station, there are certain risks of personnel security. In order to solve these problems above, an information management system of spiral welded pipe which bases on technology of PLC and barcode is designed, firstly using barcode technology into production line of steel pipe as the medium of information tracking and management. The corresponding mechanical structure of barcode printing and recognition is designed, multiple mobile scanners and fixed scanners are equipped for different stations. PLCs and PCs control movements of the scanners and mechanical structures through Ethernet, realizing information paperless recording and transmission of spiral welded pipe, greatly improving the automation level of production line.


  • Multiday Pair Trading Strategy using Copula
    Apoorva Uday Nayak, Vishwanjali Gaikwad and Megha Ugalmugale, College of Engineering Pune, India
    ABSTRACT
    Pair trading is a market neutral strategy meant to generate profit regardless of whether equities rise or fall. The strategy of matching a long position(buy) with a short position(sell) in two historically correlated stocks is known as pair trading. Some traditional pair trading techniques use correlation or cointegration as a dependence measure and assume symmetric distribution of data along the mean zero. However, the occurrence of non-linear distributions is quite frequent in financial assets and thus the use of linear correlation coefficient as a dependence measure is erroneous and may lead to misleading results. This may trigger wrong trading signals and may fail to recognise profit opportunities. This paper presents an overview of applications of different copulas to develop a model to suggest profitable strategy by analyzing data and providing trading signals using machine learning and other statistical analysis techniques. Copula is a relatively new pair trading technique. Copulas are much more realistic. They can be applied regardless of the form of marginal distribution thus providing much more robustness and flexibility in practical applications.


  • Automation in Warehouse Storage & Retrieval System
    Vikrant Pravin Kamat, Vidyavardhini's College of Engineering and Technology, India
    ABSTRACT
    Automation is the process of reducing human efforts and using machines for performing the same operations Automation in warehouse storage and retrieval system will play an important role in increasing levels of production, reducing fatigue in operators and also minimizing work load on workers. Automation will make the overall system safe, robust and highly efficient. Industrial automation tools like Programmable Logic Controllers and Supervisory Control and Data Acquisition play an important role in implementing this concept. This paper also proposes the design of an automated storage machine and an automated retrieval machine using a lead screw based vertical linear actuator and a warehouse model in which there are multiple multi-level conveyor racks for storage purposes. Also, piezoelectric cells can be installed under the system of flat bed conveyors and the EMF produced due to piezoelectric effect because of the weight of the load on the conveyors, can be used for supplying electric potential to drive the actuators involved in the process. The proposed automated storage and retrieval mechanisms can be typically used in applications where there is a very high volume of loads being moved into and out of storage, and where accuracy is important to avoid any critical damages to the load.


  • Towards Modeling Disease Outbreak Notification Systems
    Farag Azzedin, Jaweed Yazdani, Salahadin Adam and Mustafa Ghaleb, King Fahd University of Petroleum and Minerals, Saudi Arabia
    ABSTRACT
    Disease outbreak detection, monitoring and notification systems play an important role in assessing threats to public health since disease outbreaks are becoming increasingly common world-wide. There are several systems in use around the world, with coverage of national, international and global disease outbreaks. These systems use different taxonomies and classifications for the detection and prioritization of potential disease outbreaks. In this paper, we study and analyze the current disease outbreak systems. Subsequently, we extract features and functions of typical and generic disease outbreak systems. We then propose a generic model for disease outbreak notification systems. Our effort is directed towards standardizing the design process for typical disease outbreak systems.