07-11-2012, 02:35 PM
ESTIMATION OF VIRTUAL 3D IMAGES BY TEXT FOR MOBILE AUGMENTED REALITY
![Microsoft PowerPoint 2007 Document .pptx](https://seminarproject.net/images/attachtypes/ppt.gif)
ABSTRACT:
This project is based on the field of Augmented Reality, in which 3D virtual objects are integrated into a 3D real environment in real time.
MODEL-BASED visual tracking has become increasingly attractive in recent years in many domains, such as robotics and Augmented Reality (AR).
We have the proposed approach in NESTOR, a system that operates in real time on a mobile phone.
In the letter case, the system automatically recognizes the new shape before learning it .
Since the main motive of our project is educational purpose , we are using chemistry formula in which we use AR technique to retrieve their corresponding formula’s .
Letters offer various benefits for AR they lend themselves to identification and pose estimation in cases of partial blockage and moderate projective distortion.
EXISTING SYSTEM :
According to the paper “ SHAPE RECOGNITION AND POSE ESTIMATION FOR MOBILE AUGMENTED REALITY “ we use the shapes to recognize the objects which are already stored in the library , when the new shapes are recognized it will store in the library .
Here we use an concept called pose estimation in which we calculate the centroid of the image that is produced before the android camera.
Object recognition
Model-based pose estimation
The user can now attach a virtual model to be augmented to the new shape and modify its different properties, such as scale and rotation.
PROPOSED SYSTEM :
In the proposed system we use alphabets and collection of alphabets to produce 3D image
Input is given in text converted to grayscale image and recognize the pattern
By calculating the pixels we segment the image as a separate template and analyzed by OCR algorithm
Then 2D is converted to 3D image
CAHRECTER RECOGNITION:
To recognize the character that we obtained from the image segmentation template we use a specialized algorithm called as Optical Character Recognition which is in short known as OCR algorithm
The OCRchie recognition algorithm relies on a set of learned characters and their properties.
It compares the characters in the scanned image file to the characters in this learned set.
It requires that an image file with the desired characters in the desired font be created, and a text file representing the characters in this image file.
Once the learned set has been read in from the image file and its properties recognized, it can be written out to a "learn" file.
This file stores the properties of the learned characters in abbreviated form, eliminating the need for retaining the images of the learned characters, and can be read in
very quickly.
CONCLUSION
We have described Nestor, a recognition and pose estimation system for planar shapes. It performs robust recognition of shapes and maintains accurate and stable 3D registration in extreme slant angles, as well as in the cases of partial occlusion. Nestor allows planar shapes to be used for registration as flexible fiducially for AR. Nestor rectifies new shapes according to previously learned shapes and automatically assigns virtual content to them according to a letter and shape class library.