1 / 2

Abstracts Spring 2010 - Dr. Nasr

The American University in Cairo Department of Computer Science & Engineering CSCE 492 Projects' Abstracts Spring 2010 Ground Up

Download Presentation

Abstracts Spring 2010 - Dr. Nasr

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. The American University in Cairo Department of Computer Science & Engineering CSCE 492 Projects' Abstracts Spring 2010 Ground Up Augmented Reality (AR), one of the leading research topics in the field of computer science, permits the overlay of sensory stimulus onto the real environment, offering the user an enhanced perception of his surroundings. Augmented stimulus can be visual, audible or even tactile, bridging the gap between the real and virtual environments. Built on the latest surface computer hardware supporting multi-touch gestures, and coupled with cutting edge 3D graphics and AR technologies; our application "Ground Up" will allow designers to give a new dimension to their projects: Multi-touch would allow one or multiple users to collaborate on the same project, gestures would permit intuitive interaction with the designs, while AR will increase the users' immersion into his creations, by visually bringing them to our three dimensional world in the form of small scale prototypes. Our product would present a practical, usable and affordable solution based on a mixture of the latest hardware and software technologies available in the tackled fields. Such an application will offer designers a new tool which puts creativity and imagination at their fingertips. Ultimately, the way people design will be redesigned. Track Me If You Can Object tracking is an important research topic in the field of the computer vision. In recent years, the increasing demand for automated video analysis together with the advent of high quality, inexpensive video cameras and computers motivate more researchers to pay a great deal of attention to object tracking issues. However, the tracking major drawback is the vast amount of computation needed to produce a reliable result in real time. Therefore for our thesis, we will enhance and develop a template-based tracker for tracking multiple objects, which adapts to several object transformations. Nevertheless, solving the speed problems by parallelizing the algorithm and writing it using CUDA to be deployed on the GPU. We will be introducing a new modified version of the Mean Shift (MS) algorithm, which will work combined with a border matching technique, in an attempt to obtain the most accurate results. MS is one of the most accurate algorithms that have ever been used in the tracking field; however, attempting to parallelize it is challenging. Fashion Forward A simple mathematical calculation would tell us that if person spends 15 minutes of his day thinking about what to wear, by the time he reaches the age of sixty he would have spent 228 days of his life thinking about what to wear! In order to minimize that, we aim at creating a recommendation system that would help people with choices about their apparel. The application will analyze images of the clothing items provided by the user to get information about them. First it identifies the color of the item by defining the dominant color in the picture. Second it identifies the type of the item by highlighting that dominant color, which will be used in the process of deciding whether the item will be categorized as a top or a bottom piece. The user will have the option of skimming through his wardrobe by selecting the characteristics of the clothes he wants to see and the software will filter his wardrobe according to the description he specified. The application will gather information about the events the user intends to attend from the user's event calendar which has the name of the event, social group, location and time. After that it will filter the clothes according to some certain criteria, and it will plan an appropriate outfit for each event. The criteria it will use for selecting suitable outfits is weather, user preference, occasion, history and availability of each item.   Then it will display the results to the user from which he/she can select the best match. Another dimension to be explored is the potential of adding some networking aspects to the application such as the ability of the application to connect to online stores for shopping services or to be offered recommendation by other users. DyNecT The goal of this thesis is to augment an existing video chat client with the ability to separate the individual chatting from his background, then add another background instead. The separation of the background is done using various image processing algorithms. Interoperability with many chat clients is achieved through the use of a virtual webcam. The objective here is to do this in real-time. The idea was inspired from the fact that the internet is getting less and less private, thus we wanted to give the person chatting the opportunity to maintain a degree of privacy if he/she does not want to share his/her surrounding environment during video chat. The project is to be implemented such that the segmenting component can be easily used in other programs or research areas. StOKie Using the technologies of Artificial neural networks, Genetic algorithms and Fuzzy logic, we aim to build an intelligent system that can give out reliable stock trading decisions by analyzing the historical data of the stocks. Web Search Engine We will gather the best results from multiple search engines, remove the duplicates, and re-rank the results. In more details, we will take the best ten results, for example, from Google, Yahoo, Ask, then take the common results, in order to return the results that cover the search engines that the user chooses to get his results from. The new features that we will provide are that we will allow the user to choose the search engines that he wants to get the results from. He will select from a list of search engines. Moreover, we will try to provide a very user friendly interface to ease the searching process. WriteRight WriteRight is a pen that is capable of detecting some hand written characters using acceleration. The whole system consists of two major parts: acceleration detection using two3D accelerometers and signal processing using Matlab. WriteRight application can be the base for a useful tool that is capable of converting handwriting to its equivalent soft copy document to be ready to be saved, edited and shared among many computers. Features: - Acceleration detection using accelerometers - Interface with microcontroller - Memory Storage - Signal Processing

  2. The user will have the option of skimming through his wardrobe by selecting the characteristics of the clothes he wants to see and the software will filter his wardrobe according to the description he specified. The application will gather information about the events the user intends to attend from the user's event calendar which has the name of the event, social group, location and time. After that it will filter the clothes according to some certain criteria, and it will plan an appropriate outfit for each event. The criteria it will use for selecting suitable outfits is weather, user preference, occasion, history and availability of each item.   Then it will display the results to the user from which he/she can select the best match. Another dimension to be explored is the potential of adding some networking aspects to the application such as the ability of the application to connect to online stores for shopping services or to be offered recommendation by other users. DyNecT The goal of this thesis is to augment an existing video chat client with the ability to separate the individual chatting from his background, then add another background instead. The separation of the background is done using various image processing algorithms. Interoperability with many chat clients is achieved through the use of a virtual webcam. The objective here is to do this in real-time. The idea was inspired from the fact that the internet is getting less and less private, thus we wanted to give the person chatting the opportunity to maintain a degree of privacy if he/she does not want to share his/her surrounding environment during video chat. The project is to be implemented such that the segmenting component can be easily used in other programs or research areas. StOKie Using the technologies of Artificial neural networks, Genetic algorithms and Fuzzy logic, we aim to build an intelligent system that can give out reliable stock trading decisions by analyzing the historical data of the stocks. Web Search Engine We will gather the best results from multiple search engines, remove the duplicates, and re-rank the results. In more details, we will take the best ten results, for example, from Google, Yahoo, Ask, then take the common results, in order to return the results that cover the search engines that the user chooses to get his results from. The new features that we will provide are that we will allow the user to choose the search engines that he wants to get the results from. He will select from a list of search engines. Moreover, we will try to provide a very user friendly interface to ease the searching process. WriteRight WriteRight is a pen that is capable of detecting some hand written characters using acceleration. The whole system consists of two major parts: acceleration detection using two3D accelerometers and signal processing using Matlab. WriteRight application can be the base for a useful tool that is capable of converting handwriting to its equivalent soft copy document to be ready to be saved, edited and shared among many computers. Features: - Acceleration detection using accelerometers - Interface with microcontroller - Memory Storage - Signal Processing The user will have the option of skimming through his wardrobe by selecting the characteristics of the clothes he wants to see and the software will filter his wardrobe according to the description he specified. The application will gather information about the events the user intends to attend from the user's event calendar which has the name of the event, social group, location and time. After that it will filter the clothes according to some certain criteria, and it will plan an appropriate outfit for each event. The criteria it will use for selecting suitable outfits is weather, user preference, occasion, history and availability of each item.   Then it will display the results to the user from which he/she can select the best match. Another dimension to be explored is the potential of adding some networking aspects to the application such as the ability of the application to connect to online stores for shopping services or to be offered recommendation by other users. DyNecT The goal of this thesis is to augment an existing video chat client with the ability to separate the individual chatting from his background, then add another background instead. The separation of the background is done using various image processing algorithms. Interoperability with many chat clients is achieved through the use of a virtual webcam. The objective here is to do this in real-time. The idea was inspired from the fact that the internet is getting less and less private, thus we wanted to give the person chatting the opportunity to maintain a degree of privacy if he/she does not want to share his/her surrounding environment during video chat. The project is to be implemented such that the segmenting component can be easily used in other programs or research areas. StOKie Using the technologies of Artificial neural networks, Genetic algorithms and Fuzzy logic, we aim to build an intelligent system that can give out reliable stock trading decisions by analyzing the historical data of the stocks. Web Search Engine We will gather the best results from multiple search engines, remove the duplicates, and re-rank the results. In more details, we will take the best ten results, for example, from Google, Yahoo, Ask, then take the common results, in order to return the results that cover the search engines that the user chooses to get his results from. The new features that we will provide are that we will allow the user to choose the search engines that he wants to get the results from. He will select from a list of search engines. Moreover, we will try to provide a very user friendly interface to ease the searching process. WriteRight WriteRight is a pen that is capable of detecting some hand written characters using acceleration. The whole system consists of two major parts: acceleration detection using two 3D accelerometers and signal processing using Matlab. WriteRight application can be the base for a useful tool that is capable of converting handwriting to its equivalent soft copy document to be ready to be saved, edited and shared among many computers. Features: - Acceleration detection using accelerometers - Interface with microcontroller - Memory Storage - Signal Processing

More Related