Combination of extraction features based on texture and colour feature for beef and pork classification
Arif Mudi Priyatno, Fiqhri Mulianda Putra, P Cholidhazia, Lestari Ningsih
Abstract
Abstract Behaviour of traders mixing beef and pork is very detrimental to consumers, especially followers of Islam because it is related to legal or forbidden food. So, consumers must be protected from these rogue traders. However, differentiating beef and pork is not easy for ordinary people, especially if you only see from one information that is the colour or texture. In this paper, we proposed a new combination of extraction features based on texture and colour features for the classification of beef and pork. The feature of the texture is to see the local information optimally by using a local optimal-oriented pattern (LOOP) so that it can provide better texture information. The colour features that will be used are hue, saturation, and value (HSV). Texture and colour features are combined into one, so that more enrich the information used. The combination of optimal local-oriented pattern features and hue saturation value gives increased accuracy for the classification of pork and beef. The results of tests that have been done show that the success rate of calcification by using a combination of features has increased. accuracy obtained is equal to 99.16 percent, recall 100 percent and precision 98.36 percent. this shows that by utilizing the colour features and texture features can provide improved classification due to increased information that can be used to do the classification.