COMBINATION OF GLCM AND KNN CLASSIFICATION FOR CHICKEN EMBRYO DEVELOPMENT RECOGNITION

Abstract

This paper presents techniques for chicken embryo development recognition based on textural analysis using Gray Level Co-occurrence Matrix (GLCM) and image classification using k-nearest neighbors algorithm (KNN). Data employed in the experiments were collected during the eighteen days of incubation, which was divided into four stages, namely Stage 1: days 1-3, Stage 2: days 4-6, Stage 3: days 7-9, and Stage 4: days 10-18. Results of the experiments reveal that the recognition accuracy rates for Stage 1, Stage 2, Stage 3 and Stage 4 are 73.33%, 60%, 66.67% and 93.33%, respectively.

Existing System

Support Vector Machine (SVM) classifier

Proposed System

The method proposed in this paper is to detect main feature of an image employed to identify the time of conception and chicken embryo development from image dataset. Generally, main feature is texture feature that can be used to classify the image data, which will be distributed in the form of tones within an image. Therefore, change in tones within the image is important for image data classification, which can be defined from the grayscale matrix by using GLCM functions to calculate and compare the occurrence of gray levels in an image or pattern of gray levels between pixels in the image in order to utilize the texture feature in the processes of image data classification and recognition. KNN algorithm is used for image data classification. The distance between datasets is computed to measure similarity of image datasets.  

Architecture

BLOCK DIAGRAM