TechSpotlight
A snap of our future software engineers
Young IIT students working at the department's software lab (top left) and screenshots from skin detection software Photo: SHAWKAT JAMIL
IT is the youth who showed a rare feat at the very beginning of their academic career through developing a project on skin detection. Its emphasis soared when the German based journal Springer-Verlag published a detail on this project. It is always a dream for the MS, MPhil and even for PhD students to have such a publication in Springer-Verlag journal that recognises research work internationally. But this time they are not the higher tiered students, they are undergraduate students of Institute of Information Technology (IIT) of University of Dhaka (DU) whose research work was recognised by the journal. As they were the very first batch of IIT's Bachelor in Information Technology (BIT) programme, so they steal the show. The research focused on automatic skin detection of people of Indian sub-continent and they obtained huge success in this field. The seven students involved in the research are Alim Ul Gias, Md Mostafa Kamal, Hossain Muhammad Muctadir, Asif Imran, Asif Khan Shakir, Saiful Islam & Mirza Rehenuma Tabassum. They developed the skin detection project in first semester are now studying in second semester of their undergraduate. Two lecturers of IIT Md Saiful Islam and Shah Mostafa Khaled supervised the project work. They said the BIT's first semester students were assigned on several projects on 'Structured Programming' course. Three groups amongst them conducted research on image processing. It was seen that they programmed software to detect the skin of Indian sub-continent at the end of their project schedule with a great success and efficiency. The software was trained and tested using human images, some of which were locally captured, and others randomly collected from the Internet. Professor Dr. Zerina Begum, Chairperson of BIT program was saying publication in a Springer-Verlag journal is a great achievement for any researcher of science and technology. For students who have completed only one semester in their undergraduate program this is an outstanding achievement. About the skin detection softwareAs the Springer-Verlag abstract says following a title "Combinatorial Colour Space Models for Skin Detection in Sub-Continental Human Images", among different colour models HSV, HLS, YIQ, YCbCr, YUV, etc. have been most popular for skin detection. But most of the research done in the field of skin detection has been trained and tested on human images of African, Mongolian and Anglo-Saxon ethnic origins, skin colours of Indian sub-continentals have not been focused separately. In this research, a comparative study of different combinatorial skin detection algorithms has been made. For training and testing 200 images (skin and non skin) containing pictures of sub-continental male and females have been used to measure the performance of the combinatorial approaches, and considerable development in success rate with True Positive of 99.5 percent and True Negative of 93.3 percent have been observed. So accurate was their project performance, said the lecturers. They explained that the target is to achieve a high detection rate and low false positive rate, that is, skin pixels must be detected in maximum and the amount of non-skin pixels classified as skin should be minimised. Challenge
The main challenge in skin detection is to make the recognition robust to the large variations in appearance of skin that may occur, like in colour and shape, effects of occultation, intensity, colour, location of light source. Imaging noise can appear as speckles of skin like colour, and many other objects like wood, cooper and some clothes are often confused as skin. Scope
The Springer-Verlag journal describes about the scopes of the skin detection software as the combinatorial methods achieve a significantly lower false-positive and false-negative rates for skin detection. Another observation of this research is to cluster the skin colour for Indian sub-continental human images under different lighting conditions. One of the directions that future research might take is to develop better heuristic ways to integrate multi-cues, including colour, texture, spatial and shape, into an even more powerful classifier for the skin detection task. Talking about the scope, Shah Mostafa Khaled said this research's result, the skin detection software can be extended to develop applications for adult image detection, motion detection, object detection, security etc. sectors. Moreover, the internet based pornography can be prevented by using this type of software integrating it with the system. What they say
Professor Dr. Zerina Begum expressed her satisfaction about the work and appreciated the devotion, determination, and discipline of the students in doing the work. She said the work done by the young students is a fundamental one. But the software has great potential to be used in real world applications. Furthermore, it is significant that they are student of first year. So, it is huge and it indicates that if we can create spontaneous interest among them, they will do far better. The researchers are hopeful that their work will open a new door in the field of skin detection. Among the students, Alim Ul Gias and Mirza Rehenuma Tabassum said their research work was based on still image. But now they are planning to work on the video image with full-proof work so that the software can be used in real-time situation to protect pornography in the internet. Md Mostafa Kamal said they were greatly motivated by their achievement in skin detection. "We found it more challenging and interesting", he added. Further development
They said the students submitted their research paper in the First International Visual Informatics Conference, which is going to be held in Malaysia in November 2009. The research been accepted with high appreciation, and the paper will be published by Springer-Verlag the famous scientific publisher of Germany, in their Lecture Notes on Computer Science (LNCS) series. About IIT
IIT started its journey on 1985 as a computer centre and moulded into IIT on 2001. IIT started its bachelor course in 'Software Engineering' in 2008-09 sessions, which is the first initiative by a public University to open specialised undergraduate on Software Engineering. It has semester based four years Honours programme and one year Post Graduate Diploma in Information Technology programme, the latter was started on 2004. Mohd Zulfiquar Hafiz, associate professor of IIT, said the general computer science department tries to cover total aspects of computer. But in IIT, we are focused on need-based profession of the market. More precisely, we are trying to develop international standard software engineer narrowing down all the aspects of computer science. That means we are targeting to make a student capable to perform as project manager, network manager, software developer and software designer- whatever or wherever one wants to find their destination in the software industry.
For all latest news, follow The Daily Star's Google News channel.
Comments