Abstract:
Images are more ambient than words similarly the 3D models are more descriptive
than 2D pictures. 3D model of any product will give the enhanced information about
it. The main objective of generating the 3D model is to reduce the amount of time
by automating the process of generating the 3D model. Augmented Reality (AR)
technologies have been recognized as offering great benefit to almost every field
and organizations, although they are not easy to implement. AR provides enhanced
information of the product by overlaying the 3D model into the real world. 3D
models can express the shape and dimensions of the products with greater precision
and they are being extensively used these days.
In this project, we have implemented a 3D Modeling application which will take
only a few 2D images as input from any image capturing device taken from different
angles and it will automatically convert it into a 3D model. The first step towards
generating the 3D model from the images is to establish point correspondence which
includes feature extraction and feature matching. Then refining correspondence is
done followed by de-calibration of 3D structure and camera pose. The final step is
texture mapping and rendering.
We are using an e-commerce portal as an example to display the 3D models
generated by our 3D Modeling application. Most often customers shopping on
e-commerce websites do not get complete information about the product by only
viewing the 2D images. They can get a better understanding of the product by using
AR View. The adaptation of Augmented Reality (AR) will take e-commerce to the
next level.
