Rancang Bangun Sistem Identifikasi Kesegaran Ikan Berdasarkan Citra Mata Menggunakan Support Vector Machine

Sari

Fish is one of the staple foods that is in great demand, especially in Indonesia because it has perfect protein content, this
content is obtained from good quality or fresh fish. The nutritional composition of fish consists of 18.13% protein, 1.90% fa t, 1.03% ash
and 78% water. Fish is a type of staple food that is easy to process without any special handling. The quality and selling value of fish is
highly dependent on the quality of the freshness of the fish. A good selection of fish is to choose the type of fish that is still fresh. The
quality of fish that is suitable for consumption can be seen in the color of the fish skin, fish eyes, gills and meat texture. However, there
are still many consumers who do not understand how to choose the quality of the freshness of the fish. so that it has an impact on health
even if consuming rotten fish can cause poisoning to digestive disorders. Based on these problems, a solution that can be app lied to the
selection of fresh fish was taken, namely by designing a Fish Freshness Identification System Based on Eye Image. The implementation of
this system is done by taking pictures of fish in jpeg format, then cropping and resizing to get the pixel size. after that the calculation is
carried out through the conversion process of cropping fish heads from grayscale images into binary images. The process to get the value,
starting with the extraction of the RGB image to form a vector value. The result of feature extraction is then classified using the Support
Vector Machine method. This study aims to identify the freshness of fish based on eye image. so that it can make it easier for users to
distinguish the freshness of fish. In testing this system, it produces an overall accuracy rate of 73%. Based on these results, it can be
concluded that the Support Vector Machine algorithm has succeeded in identifying the freshness of fish.

Keywords — Support Vector Machine, RGB, Identifikas, Citra Mata