Beef Image Classification using K-Nearest Neighbor Algorithm for Identification Quality and Freshness

S. Agustin, R. Dijaya

Research output: Contribution to journalConference articlepeer-review

11 Citations (Scopus)

Abstract

Nowadays many traders are cheating in selling meat that is not feasible for consumption to get a greater profit by mixing good meat with meat that is not feasible for consumption. So cause anxiety for the community because it has dangerous content. One way to help people in selecting meat through image processing can recognize objects. The purpose of this research is to know the quality of beef which is good and feasible to be consumed using the co-occurrence matrix to classify the meat image with K-NN algorithm. This research can be utilized to be able to distinguish the types of meat based on colour and texture. The data used in this research are 60 image data consisting of 30 images of fresh meat and 30 images of rotten meat. The classification process uses test data in order to distinguish the types of fresh meat and rotten meat. The mean feature value of this method with the feature features of the highest feature extraction is the homogeneity value feature. The results showed that the performance of the system using the KNN method to identify the quality of meat based on colour and texture can detect the type of beef.

Original languageEnglish
Article number012184
JournalJournal of Physics: Conference Series
Volume1179
Issue number1
DOIs
Publication statusPublished - 30 Aug 2019
Externally publishedYes
Event1st International Conference on Computer, Science, Engineering and Technology, ICComSET 2018 - Tasikmalaya, West Java, Indonesia
Duration: 27 Nov 201828 Nov 2018

Fingerprint

Dive into the research topics of 'Beef Image Classification using K-Nearest Neighbor Algorithm for Identification Quality and Freshness'. Together they form a unique fingerprint.

Cite this