**PPT On Identification and classification of similar looking food grains
**
Download
**Food Grains Presentation Transcript:**
1.Identification and classification of similar looking food grains

2.Problem Statement

Comparative study of ANN and SVM classifier models

taking a case study of identification and classification of

similar looking food grains is the main aim of this work.

The problem is to process images of selected type of grains and extract the features from the samples based on RGB, HSV and wavelet based texture.

Developing the different classification models using various set of features.

Identification of similar looking grains and test their performance based on their rate of recognition.

3.Methodology for Classification of Similar Looking Grains

4.Images of Grains Samples

5.Feature Extraction

1.Color Features

RGB (Red, Green, and Blue) Color Model

Extraction of RGB features is separation of RGB components from the original Color image sample.

After separation the features viz. Mean, Standard Deviation, Variance and Range are computed.

HSV (Hue, Saturation and Value) Color Model

HSV- used to distinguish one color from another.

Hue is the angle measured from the red axis to the point of interest .

Saturation refers to relative purity or the amount of white light mixed with a hue.

Value or Brightness embodies the chromatic notion of intensity.

2.Wavelet Texture Features

Texture is a connected set of pixels that occur repeatedly in an image .

It provides the information about the variation in the intensity of a surface by quantifying

properties such as smoothness, coarseness, and regularity.

Texture feature extraction is carried out by decomposing an image using discrete wavelet transfom

6.Algorithm for Color Feature Extraction

Input: Original 24-bit color image.

Output: 18 color features

Start

Step 1: Separate the RGB components from the original 24-bit input color image.

Step 2: Obtain the HSV components from RGB components using following equations.

7.Color and HSV Features with Values for Mustard Sample

8.Multi-level Wavelet Decomposition

Wavelet is a mathematical function used to divide a given function or continuous-time signal into different scale components.

Approximated image LL is obtained by low pass filtering in both row and column directions.

The detailed images, LH, HL, and HH, contain high frequency components

9.Algorithm for Wavelet-based Textural Feature Extraction

Input: Original color image.

Output: 42 features.

Start

Step 1: Convert the original color image to gray image.

Step 2: Calculate the level-1 Wavelet transform and hence decompose the signal into

LL, (low frequency components) and LH, HL and HH (High frequency

components in horizontal, vertical and diagonal)

Step 3: Consider vertical details only.

Step 4: Find the Mean, Variance, Range, Energy, Homogeneity, Maximum Probability,

Inverse Difference Moment (IDM) .

Step 5: Copy the values obtained to a feature vector

Step 6: Repeat the steps 4 and step 5 for remaining detailed coefficients set namely,

horizontal and diagonal coefficients.

Step 7: Repeat the steps 4 to 6 for third level decomposition and combine the feature set

into a single feature vector.

Stop.

10.Texture Descriptors