Shapes Classification using SVM Test phase, srs, design phase and source code final deliverable
Project Domain/Category
Image Processing
Abstract / Introduction
Image processing is a method of performing operations on an image with the intention to enhance it or acquire useful information from it. Image processing has numerous programs in fields including medical, protection, industry, remote sensing, sample reputation, and video processing, among many others. In this mission we are able to create an software that should pick out one-of-a-kind items like rectangle, square and circle etc., based on their capabilities from the input image.
Functional Requirements:
1. For this utility you have to use four shapes i.E., square, circle, rectangle and celebrity.
2. Use 10 one of a kind sizes for every of the form.
Three. Create 10 snap shots of each of the shape and size the use of jpeg format in MS paint or some other software program.
4. All the photos ought to have identical peak and width of 128 X 128 pixels.
5. All the photographs must have black and white. Black form with white heritage.
6. The entire dataset should consist of four hundred snap shots.
7. Divide the dataset into 70% schooling set to train a version and 30% checking out set to check the skilled version. The division must now not be biased.
8. You need to use Support Vector Machine (SVM) for the type.
9. Extract capabilities from the pix the use of exclusive integrated features of MATLAB like regionprops() and many others.
10. Train SVM version the use of the schooling dataset.
11. Test SVM model using the take a look at dataset.
12. The device need to have the functionality to check the whole trying out set after which generate the accuracy report for all of the shapes.
Thirteen. The system need to also have the capability to check a unmarried image.
14. You ought to create a right interface for most of these activities.
15. You must use one-of-a-kind built-in capabilities of MATLAB wherein relevant.
Tools & Technologies:
Preferred device and generation: MATLAB (Any trendy model of MATLAB)