Abstrak- Pelayanan prima yang cepat dan tepat harus diberikan oleh pemerintah desa kepada masyarakat. Namun kesan layanan birokrasi pemerintahan berbelit-belit sudah menjadi pemikiran kebanyakan masyarakat. Hal ini didasari atas fakta yang masyarakat alami sendiri ketika mereka memanfaatkan layanan di kantor desa. Sistem layanan mandiri dapat menjadi solusi untuk mengurangi jalur birokrasi yang dianggap berbelit-belit tersebut tanpa harus menghilangkan peran pihak-pihak terkait. Pada sistem layanan mandiri ini masyarakat dapat mengakses layanan seperti mengurus surat keterangan dan surat pengantar untuk berbagai keperluan. Untuk meminimalkan proses pengetikan oleh masyarakat digunakan image processing untuk mendapatkan data Nomor Induk Kepenedudukan (NIK) pada KTP. Image processing dilakukan dengan cara bileveling, smearing, penentuan baris dan ekstrasi citra digital ke citra biner yang menghasilkan vektor. Pola vektor tersebut di uji dengan jaringan syaraf tiruan metode backpropagation yang telah dilatih menggunakan 61 unit vektor data sampel. Hasilnya jaringan syaraf tiruan mampu mengenali (NIK) pada KTP, NIK tersebut dijadikan input pada system layanan mandiri. Kata kunci: Pengolahan citra, Run Lenght Smearing Algorithm, Jaringan Syafat tiruan, Backpropation Abstract- Excellent service should be provided promptly by the village government to the community. However, the impression of bureaucratic government service has become the most people thought. This is based on the facts that people themselves experience when they make use of the services in the village office. Self-service system can be a solution to reduce the bureaucratic path that is considered convoluted without having to eliminate the role of the parties concerned. In this self-service system the community can access services such as taking care of certificates and cover letters for various purposes. To minimize typing activities by community used image processing to obtain data population identification number on the ID card. Image processing is done by bileveling, smearing, line determination and extracting digital images into binary images that generate vectors. The vector pattern was tested with artificial neural network backpropagation method which has been trained using 61 vector units sample data. The result is artificial neural networks are able to recognize data population identification number on the ID card. The population identification number is used as inputs on the system self-service. Keywords: Image processing, Run Lenght Smearing Algorithm, Articial neural network, Backpropagation