Bilkent University - Computer Engineering Department

CS491-492: Senior Design Project I and II: 2017-2018

Projects supervised by Ibrahim Korpeoglu
[CS491 Page] [CS492 Page] [Piazza] [CS Fair 2017] [Current Details]

Schedule:

Schedule Calendar

Project 1: AndLit

Members:

  • Erin Avllazagaj, 21402962, erin.avllazagaj AT ug.bilkent.edu.tr
  • Xheni Caka, 21403083, xhenicaka AT gmail.com
  • Gunduz Huseynli, 21402821, gunduzhuseynli12 AT gmail.com
  • Mehmet Furkan Dogan, 21502274, mehmetfdogan AT gmail.com
  • Muhammad Hamza Khan, 21402885, hamza.khan AT ug.bilkent.edu.tr

Jury Members: Ibrahim Korpeoglu (supervisor), Cigdem Gunduz Demir, Ozgur Ulusoy.

Innovation Expert: Mustafa Sakalsiz, T2.

Links: [WebPage] [GitHub]

Description:AndLit is a project which aims creating a smart earpiece which is going to have various functions based on image processing related features. The main feature which AndLit is going to have is face recognition. By taking a picture from the button in the earpiece the user will be able to recognize people's faces. If the people in front of the user is not in the database then it is saved as e new person with a name given by the user using the mic positioned in the earpiece. If the person is already in the database then the user will be listening to his/her name in his ears from the earpiece. If there are more then one person in the taken picture then the person in the middle will be selected to be processed. The face recognition process is handled in a smartphone connected to the earpiece by wireless. Also data analysis and classification is done in the smartphone. Text to speech is another feature which allows users to turn text images to voice just by taking a picture of the text. This way the users can read different stuff like books but even other texts which are readable. An interesting ability of AndLit is that it can synchronize with the social media accounts that the user possesses. This way the search is not done only through the people who are saved from the user but also for people who have social media accounts. It goes even further allowing users to share their data in a pool/network giving them the opportunity to know people who have shared their data too with the cost of being shared as well so you don't need a fresh start, just join a pool.

In the News


Project 2: Bumpster

Members:

  • Gülsüm Güdükbay, 21401148, gudukbay AT gmail.com
  • Duygu Durmuş, 21401178, duygu.durmus AT ug.bilkent.edu.tr
  • Doğukan Yiğit Polat, 21401797, yigit.polat AT ug.bilkent.edu.tr
  • Bora Bardük, 21401029, bora.barduk AT ug.bilkent.edu.tr
  • Muhammed Çavuşoğlu, 21400653, m.cavusoglu AT ug.bilkent.edu.tr

Jury Members: Ibrahim Korpeoglu (supervisor), Bulent Ozguc, Ugur Gudukbay.

Innovation Expert: Armagan Yavuz, Taleworlds.

Links: [WebPage] [GitHub]

Description: The system will generate bump and physically based rendering maps using a specifically trained neural network. It will be able to identify the topological properties of the surface, as well as the material properties using a single photograph. It performs the operations in the server and communicates with the front-end applications on various platforms. It will provide rather large benefits for the industry, as the previous solution to this kind of topographical analysis required expensive methods which require personnel with expertise, expensive equipment, and time. Since there are no other tools that extract physically based rendering (bump, metalness, roughness, gloss or ambient) maps with one or more photos using a neural network, this application will be beneficial in terms of accuracy and usability. After the extraction of the user selected mapping, the user will be able to obtain the resulting map and fit it onto the original photo to see the material properties or bump to create a realistic image.


Expectations
  • I expect teams to start making their hands dirty as soon as possible. They should not wait till the second semester. They should start playig with tools, writing test code, etc., in the first semester, right after submitting their proposal. Each team should also spend a good amount of time with discussions and brainstorming sessions.
  • The teams should report their progress regularly to the instructor.
  • Each document needs to be written and checked very carefully..
  • Each team member should participate in every aspect of the project (specification, analysis, design, documentation, coding)
  • If you are using a library, subsystem or tool as part of your project, make sure the interesting thing in the project is not only done by that code. Your code needs to include interesting and involved features and operations built on top of the available software/hardware. For example, just adding GUI to an exsisting software which is already implementing a lot of interesting features and algorithms is not enough.
  • You will do a serious demo of your project both at your jury meeting and at CS Fair. You will prepare a presentation for the jury meeting. Each member needs to present.
  • The project should require serious engineering design and coding, so that 5 hard-working engineers can finish the project only in 8 months, not in 7 months or less. To give an example for the amount of code to be written, each student needs to write 10000 lines of code, hence the projects need to involve 50000 lines of new code, at least.