Integrated System Design for Broadcast Program Infringement Detection

Supervision of television and radio broadcast programs by the Komisi Penyiaran Indonesia Jawa Tengah is still being carried out manually i.e. direct supervision by humans. The process certainly has some weaknesses related to human error, such as tiredness and weary eyes. Therefore, we need intelligent software that could automatically detect broadcast infringement. Up to now, research in this area has not been carried out. This research was aimed at designing an integrated system to detect broadcast infringement, including architecture design and main module interface design. Two main stages in this system are Indonesian-language speech recognition and detection of infringements of the broadcast program. With the method of Mel Frequency Cepstral Coefficients (MFCC) and Hidden Markov Model (HMM), a speech recognition application that uses 1050 sample data produces about 70% accuracy rate. This research will continue to implement the design created using speech recognition applications that have been developed.