DETEKSI SITUS PORNOGRAFI BERDASARKAN GAMBAR MENGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK

Ahmad Hunaepi (Sekolah Tinggi Manajemen Informatika dan Komputer (STMIK) Eresha)
Makhsun Makhsun (Sekolah Tinggi Manajemen Informatika dan Komputer (STMIK) Eresha)
Sarwani Sarwani (Universitas Pamulang)

TEKNIK INFORMATIKA
Vol 12, No 2 (2019): JURNAL TEKNIK INFORMATIKA

Today, the development of information technology is very fast. It is also followed by the growth of internet sites that are very fast, no exception the growth of sites contains pornographic. The Association of Indonesian Internet Service Providers (APJII), in 2018, reports as many as 55.9% of Indonesian internet users had experienced of appearance pornographic content suddenly. This study aims to detect pornographic sites based on images using the convolutional neural network method. To find out the level of accuracy in detection of pornographic sites, a training process with a dataset of 300 images is required. From the test results, an accuracy rate of 85% in the detection of pornographic sites is obtained.