International Journal of Computer Science and Engineering

1-37

Authors: FURKAN GÖZÜKARA & SELMA AYŞE ÖZEL

Nowadays, the importance of focused web crawlers is more than any time before. As the web has become massive and spam my, it is now essential to have focused web crawlers that can crawl only the targeted websites and obtain the necessary information. Instead of relying on the available public general web crawlers, today, developing a focused web crawler for the targeted web pages is preferred to increase success of information retrieval. In this paper, the challenges encountered and the proposed solutions to attempt these problems are presented, while developing an original hand-crafted, full scale, robust and effective focused web crawler for E-commerce sites, named as EcCrawler, which is developed in C# programming language by using .NET 4.5 framework and MS-SQL Server 2014 database management system. Most of the crawling challenges have been discussed before in the literature, however in this paper, practical implementation and .NET framework based solutions that includes thread pool initialization, exception handling, task parallelism, HTTP compression, duplicate web page resolution, number of concurrent connections to the same host, database communication, resource sharing between threads, etc. are presented and the proposed solutions are empirically evaluated. The experimental evaluation shows that applying the proposed solutions improve EcCrawler’s crawling speed over 400% and UI responsiveness over 100%. The proposed solutions may be applicable to any software that is developed by using .NET framework.

43-50

Authors: P. DINESH KUMAR & B. ROSILINE JEETHA

Computer vision is one among the thrust research area in the field of Image processing. Facial expression recognition is one among the thrust research dimension in computer vision. The process of recognition and identification is important due to the similarity of facial expressions. The motivation behind this research area is its capability to resolve an image processing problem and its wide range of applications. The main rationale of all images processing and computer vision algorithms is to build the visual data in a useful manner. For this reason in the domain of computer vision, the facial expression recognition begins with its applications in the HCI (Human Computer Interaction) where visual look of human, sight and touch sensations (also known as moods) and voice are utilized at the same time. This research article reviews several literatures pertaining to facial expression recognition.

67-78

Authors: MIHIR. N. DUDHREJIA, CHETAN B. BHATT & MEHUL K. SHAH

The significance of measurement of grain quality has been felt since way back to a century old. However it is tedious, but very important measurement is to measure the individual kernel’s qualitative analysis. Analyzing the grain sample manually is more time consuming and complicated process, and having more chances of errors with subjectivity of human perception. In order to achieve uniform standard quality and precision, machine based techniques are evolved, solely on its prime advantage of reproducing the same qualitative result efficiency again and again. Recent developments in the field of image processing, has opened up wide scope of its use for sample analysis too. Various applications of Image processing are seen in the field of agriculture, biomedical engineering, food and drug industry and many others. Food application mainly caters the qualitative aspect of various food and dairy products. In this article, an attempt is made to investigate techniques used for the quality analysis. The main attempt is to compare relative applicability of human v/s machine based approach of analysis. Machine based techniques can be further classified as, offline grain analysis technique and online grain analysis technique. Both techniques are having their own limitations. Offline techniques consumes more time for sample preparation. On the other hand, online techniques suffer from less processing speed and kernel missing part while processing. Research gaps are identified with respect to the both techniques’ limitations and new intelligent and accurate grain analyzer technology is evolved to enhance speed and accuracy by removing deficiency of the existing systems. Moreover offline and online grain image analyser features can be combined and enhanced to prepare a fully automated grain analyser to deal with different kind of grain varieties.

89-96

Authors: MOHAMMED JAWAR KHAMI

When using internet as main communication infrastructure, people apart need their information to be protected from other third parties. Two widely techniques are used for it, cryptography and steganography. In cryptography the existence of the encrypted message is visible to the world. While steganography conceals the very existence of the message. This paper concerns with steganography. It deals with text-in-text data hiding technique. The method of using non-printable characters of the Unicode standard characters are chosen to encode and hide English text into another English cover text to produce a stego text that can be used later, at the other end of the communication media, to extract and recover the exact secret text. Two text-in-text and text-from-text hiding and extraction algorithms are written and coded in matlab programming language. Merits and drawbacks are shown for the proposed algorithms. The designed algorithms result in adding many enhancements to the implementation of the basic method of using the non-printable characters of the Unicode standard characters to text-in-text hiding technique. These enhancements include the size minimization of stego text file, reducing of hiding and extraction processing time and thus reducing communication time. From security point of view, key has been used to encrypt the secret text before hiding it and decrypt the text at the extraction stage.