Projector Management Using Hand Gestures

There are very few interfacing technologies to manipulate real-time behavior for operating computer systems. During a presentation, it is inconvenient to control the computer and simultaneously explain the topic. There are several alternatives like speech recognition, laser-based gloves but these te

Project Title

Projector Management Using Hand Gestures

Project Area of Specialization

Artificial Intelligence

Project Summary

There are very few interfacing technologies to manipulate real-time behavior for operating computer systems. During a presentation, it is inconvenient to control the computer and simultaneously explain the topic. There are several alternatives like speech recognition, laser-based gloves but these techniques are less portable and have less accuracy. Our system, interactive wall display, allows you to use hand gestures to control OS on the projected screen. We have developed an effective human-computer interface for a virtual mouse system in a projector-camera configuration without any hardware requirement. To experience our project camera source is required to detect the user's hand.

It is implemented in a camera-projector system to achieve an augmented reality tool. In this configuration the main problem is that the hand surface reflects the projected background, thus we apply a robust hand segmentation method. Hand localizing is based on a background subtraction method, which adapts to the changes of the projected background. Hand poses are described by a method based on modified Fourier descriptors, which involves distance metric for the nearest neighbor classification. The proposed classification method is compared to other feature extraction methods. We also conducted tests on several users.

Previous system had used data gloves or markers for input in the system. We have no such constraint for using the system. The user can give hand gesture in view of camera naturally.

Project Objectives

The aim of camera and projector-based configurations is that the user interaction should be performed with the projected image instead of applying computer interfaces indirectly. A projector-camera pair is used to display the user interface on the projected surface where the camera acquires (camera image) the projected information (projected image) and the gestures of the user provide feedback about the interaction. Our goal is the development of a more natural interface in a camera-projector system using hand gesture analysis. The proposed system detects hand poses by shape analysis resulting in a larger vocabulary set in the communication. The correct shape detection is solved in the presence of a cluttered background.

Project Implementation Method

In order to extract features and recognize gesture following method is proposed.

1. Researching and gathering all the essential tools that will be required in building the whole system. Most probably we’ll work on software and camera implementation. 
2. Researching the code and understanding the different ways we can implement the code to make system recognize the hand gestures.
3. Researching different algorithms and making a library of multiple hand gestures that system can understand.
4. Increasing the Efficiency of the whole system to make to recognize the gestures in low light conditions by applying the different algorithms and checking which suits the best and working on other bug fixes and improving the overall design of the whole system.

Benefits of the Project

1. It is easy to design and manufacture as all the components are easily available.
2. It is portable and hence can be placed anywhere.
3. This system can be used for teaching purpose as teacher can teach through hand gestures.
4. It can improve the interaction of Human Computer Interaction.
5. It can be implemented in softboards and used as advertisment purposes.

Technical Details of Final Deliverable

The vision-based gesture recognition system and the camera-projector configuration form a natural way to control multimedia presentations or manipulate directly the projected image. Our hand pose recognition system offers more freedom in communication for the speaker if compared to other methods using camera-projector systems. The above work has shown that the modified Fourier-based method is robust even with a small training set, so the modification of gesture vocabulary or retraining of gestures is more efficient. We have tested our feature extraction method against several methods from the literature and showed that our method is significantly efficient. We have shown that measuring the detection results by maximum likelihood approach in the time domain could significantly improve the recognition efficiency. Gesture-based systems are considered as typical user-independent recognition tools. Users would like to use them with high recognition efficiency without preliminary training of gestures. In our consecutive work we deal with interactive training of gestures to avoid retraining of all gestures if an untrained user would like to use the system.

Final Deliverable of the Project

Hardware System

Core Industry

Education

Other Industries

IT , Media

Core Technology

Artificial Intelligence(AI)

Other Technologies

Internet of Things (IoT)

Sustainable Development Goals

Quality Education, Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com