Volume: 12
Number: 2
December 2015

To see an article directly, click its Title. To see abstract, click its [Abstract] link.

• Hand Gesture Recognition using Depth Information and Dynamic Time Warping
  Aboubakar Mountapmbeme, Hasan Mahmud, and Md. Kamrul Hasan
  Abstract PDF
  Hand Gesture Recognition using Depth Information and Dynamic Time Warping
  Aboubakar Mountapmbeme, Hasan Mahmud, and Md. Kamrul Hasan

  Abstract

  Web-based learning (WBL) is a growing demand in most institutions in different countries in the world. It offers independent and student centered learning to students through which they can work and to explore learning materials at their own pace when they want. They study more deeply in their areas of interest. Although WBL has the potential benefits of widening educational opportunities, many institutions in Uganda as one of the developing countries, are still struggling to integrate and gain benefits from it in their teaching and learning systems due to certain difficulties. The main purpose of this study is to find out the related possible barriers to teachers in implementing WBL in Uganda basing on the reviews and selected documents. After carefully investigating previous literature and related documents, the authors identified difficulties that are: infrastructure related, institutional related, individual teacher related and other problems. Identifying the difficulties to its implementations has impact on finding out possible measures to overcome these difficulties and to formulate related policies on implementing and designing web based learning in Uganda.
• An Efficient Circular Block Approach for Copy-Move Forgery Detection
  Md Sirajus Salekin, Rafsanjany Kushol, and Md. Hasanul Kabir
  Abstract PDF
  An Efficient Circular Block Approach for Copy-Move Forgery Detection
  Md Sirajus Salekin, Rafsanjany Kushol, and Md. Hasanul Kabir

  Abstract

  Copy-move forgery is a special type of digital forgery in which a part of the image is copied and pasted somewhere else in the same image with the intent to cover any important image feature. In image forensics, to detect this type of forgery we need a robust detection method which ensures the correct detection even if the image is noisy, compressed, blurred, scaled, rotated, flipped, etc. In this paper, we are interested in the detection of copy-move forgery more accurately especially in case of noise, blur and compression by reducing the number of false positives and negatives. To address this issue, we have developed an efficient overlapping circular block approach for detecting the copy-move forgery using the mean and contrast information of the overlapping blocks of the image. The circular block approach ensures rotation invariance and mean-contrast feature selection approach and comparison ascertain better performance especially for compressed image with low quality factor and also for the blurred image. The proposed approach has been evaluated and compared with competitive approaches using comprehensive image dataset. Our experimental results indicate that our method can detect duplicated regions in copy-move forgery with higher accuracy, especially for JPEG compressed, blurred and noisy images.
• Constraints Encountered by Teachers to the Implementation of Web-based Learning in Uganda
  Baguma Asuman, Md. Shahadat Hossain Khan, and Che Kum Clement
  Abstract PDF
  Constraints Encountered by Teachers to the Implementation of Web-based Learning in Uganda
  Baguma Asuman, Md. Shahadat Hossain Khan, and Che Kum Clement

  Abstract

  Web-based learning (WBL) is a growing demand in most institutions in different countries in the world. It offers independent and student centered learning to students through which they can work and to explore learning materials at their own pace when they want. They study more deeply in their areas of interest. Although WBL has the potential benefits of widening educational opportunities, many institutions in Uganda as one of the developing countries, are still struggling to integrate and gain benefits from it in their teaching and learning systems due to certain difficulties. The main purpose of this study is to find out the related possible barriers to teachers in implementing WBL in Uganda basing on the reviews and selected documents. After carefully investigating previous literature and related documents, the authors identified difficulties that are: infrastructure related, institutional related, individual teacher related and other problems. Identifying the difficulties to its implementations has impact on finding out possible measures to overcome these difficulties and to formulate related policies on implementing and designing web based learning in Uganda.
• An Enhanced Local Texture Pattern for Effective Face Feature Description
  Md. Moniruzzaman, Shaikh Jeeshan Kabeer, and Muhammad Mahbub Alam
  Abstract PDF
  An Enhanced Local Texture Pattern for Effective Face Feature Description
  Md. Moniruzzaman, Shaikh Jeeshan Kabeer, and Muhammad Mahbub Alam

  Abstract

  Feature representation technique for facial images is of paramount importance in fields ranging from biometrics and HCI to digital media, gaming and security. Researchers are focusing on new ways to perform automatic recognition of facial expression primarily because the human face is capable of ex- pressing an amazing range of feelings and emotions. Human expression detection is challenging because of variation of images in intensity, pose, lighting, occlusions etc. In this paper, a new micro pattern based feature representation technique is proposed for facial images, referred to as Angular Response Pattern (ARP). The proposed method applies directional Sobel masks on each pixel in the original image and generates edge responses. Pairs of the response values are used to calculate the angle information and based on this angle values ARP code is generated for each pixel. The image is divided into a number of sum-images and a histogram is generated for each sub image. Each bin of the histogram is updated with the magnitude information of the corresponding pixel. The feature vector is generated by concatenating all the histogram. The final histogram is then further smoothed by applying a square root operation. The proposed ARP method has been applied on publicly available benchmarked image databases for performance evaluation. Performance analysis with state-of-the-art approaches shows that the proposed method outperforms them all. Thus our approach ex- hibits robustness against illumination variations and random noise. This approach can further be applied towards applications such as person identification, age estimation, gender classification.
• Connecting the Disconnected: A Combination of DTN, CDN and TCP/IP Approach
  Ashraful Alam Khan, K. Habibul Kabir, and Muhammad Mahbub Alam
  Abstract PDF
  Connecting the Disconnected: A Combination of DTN, CDN and TCP/IP Approach
  Ashraful Alam Khan, K. Habibul Kabir, and Muhammad Mahbub Alam

  Abstract

  The information technology plays a vital role in the development of every aspect of our lives. Internet, one of these many technologies, is the most utilized entity for accessing the ocean of information. But, many remote regions of the world are still de- prived of these information available today. The prime reasons behind this deprivation being inadequate in- frastructure facilities and remoteness of the concerned places. But these could be overcome by implementing Delay Tolerant Networks (DTNs), which is a supporting communication in a rather unorthodox way for the intermittent networks, where direct link at all times may not be available. This paper proposed a practical implementation of a network architecture by combining Delay Tolerant Networks (DTN), Content Delivery Network (CDN) using TCP/IP with the help of wireless connectivity enabled devices. The network scheme aim to be implemented in the third world countries where sometimes technological amenities are scarce.