27-04-2011, 09:23 AM
Presented by
PUSHPAVALLI R
[attachment=12886]
Objective:
To design and develop GestureCam using FPGA based smart camera so as to enable gesture controlled Web Navigation.
The long term goal of GestureCam and GestureBrowser is to develop small, self contained and intelligent device that can be easily embedded into HCI environment and can recognize user’s gestural behavior.
Introduction:
Smart camera, or an intelligent camera is an embedded system which captures and process image to extract application specific information in real time.
Smart cameras can have may applications such as Automatic control systems, Video surveillance, security, machine vision systems, human computer interface.
Unlike UI technologies which are computer centered, an Multi Modal User Interface MMUI is user centered and allows a user to communicate with computer using his or her natural communication modalities.
Gesture Recognition is an important part of MMUI systems.
A GestureCam is a smart camera which can perform simple head and hand
gesture recognition and can be used in many desktop MMUI applications.
GestureCam is a smart camera built from scratch, that is not based on a
commercial camera which provides processed analog or digital outputs.
Gesture Browser is an extension to the Mozilla Firefox browser which uses the
GestureCam to capture and recognize a user’s head and hand gestures to
control web navigation.
Smart Camera Design Process :
The steps that are followed in the design of GestureCam are:
Step one: Application Requirements Specification.
Step two: System Architectural Design.
Step three : Proof of Concept
Step four : Algorithmic Conversion
Step five : Integration and Debugging
Step six : Test and Evaluation
GestureCam and GestureBrowser Proof of Concept:
The main steps involved in vision based gesture recognition include these stages:
Image preprocessing: To perform color interpolation and to improve signal
to noise ratio.
Object Segmentation: To localize and segment objects of interests, such as
head, face and hands.
Feature Extraction: To extract a small set of salient parameters to represent
each gesture and provide good distinguishability between users.
Gesture Classification: A pattern recognition task which compares incoming
feature vectors against those from a database of predefined gesture
representations.
GestureCam Design and Development:
The idea of GestureCam is to design and develop a standalone, FPGA based smart camera that can perform all the processing tasks as shown in figure.
The Design and Development of GestureCam mainly involves:
System Architecture
Algorithm Design
Algorithms conversion and implementation on FPGA
System Architecture:
GestureCam consists of three parts : an Image capture unit(ICU), an FPGA based gesture recognition unit (GRU), and a Host and display unit (HDU) as shown in figure.
Algorithm Design:
Figure shows all functional modules performed by the FPGA. The camera control module receives configuration parameters from the host PC to adjust behavior and performance of the image sensor and various processing modules.
PUSHPAVALLI R
[attachment=12886]
Objective:
To design and develop GestureCam using FPGA based smart camera so as to enable gesture controlled Web Navigation.
The long term goal of GestureCam and GestureBrowser is to develop small, self contained and intelligent device that can be easily embedded into HCI environment and can recognize user’s gestural behavior.
Introduction:
Smart camera, or an intelligent camera is an embedded system which captures and process image to extract application specific information in real time.
Smart cameras can have may applications such as Automatic control systems, Video surveillance, security, machine vision systems, human computer interface.
Unlike UI technologies which are computer centered, an Multi Modal User Interface MMUI is user centered and allows a user to communicate with computer using his or her natural communication modalities.
Gesture Recognition is an important part of MMUI systems.
A GestureCam is a smart camera which can perform simple head and hand
gesture recognition and can be used in many desktop MMUI applications.
GestureCam is a smart camera built from scratch, that is not based on a
commercial camera which provides processed analog or digital outputs.
Gesture Browser is an extension to the Mozilla Firefox browser which uses the
GestureCam to capture and recognize a user’s head and hand gestures to
control web navigation.
Smart Camera Design Process :
The steps that are followed in the design of GestureCam are:
Step one: Application Requirements Specification.
Step two: System Architectural Design.
Step three : Proof of Concept
Step four : Algorithmic Conversion
Step five : Integration and Debugging
Step six : Test and Evaluation
GestureCam and GestureBrowser Proof of Concept:
The main steps involved in vision based gesture recognition include these stages:
Image preprocessing: To perform color interpolation and to improve signal
to noise ratio.
Object Segmentation: To localize and segment objects of interests, such as
head, face and hands.
Feature Extraction: To extract a small set of salient parameters to represent
each gesture and provide good distinguishability between users.
Gesture Classification: A pattern recognition task which compares incoming
feature vectors against those from a database of predefined gesture
representations.
GestureCam Design and Development:
The idea of GestureCam is to design and develop a standalone, FPGA based smart camera that can perform all the processing tasks as shown in figure.
The Design and Development of GestureCam mainly involves:
System Architecture
Algorithm Design
Algorithms conversion and implementation on FPGA
System Architecture:
GestureCam consists of three parts : an Image capture unit(ICU), an FPGA based gesture recognition unit (GRU), and a Host and display unit (HDU) as shown in figure.
Algorithm Design:
Figure shows all functional modules performed by the FPGA. The camera control module receives configuration parameters from the host PC to adjust behavior and performance of the image sensor and various processing modules.