Skripsi
IMPLEMENTASI METODE CENTER SCAN UNTUK IDENTIFIKASI POLA EKSPRESI WAJAH PADA ROBOT PENGANTAR OBAT
Face identification is a form of Machine Learning application, specifically using Deep Learning, that is commonly encountered in our daily lives. In its implementation, there are many algorithms used, and those employed in image processing for embedded schemes must be able to operate effectively on the intended hardware system. One example of such an algorithm is the Weightless Neural Network (WNN). This study aims to detect static faces using the Haar-Cascade Classifier method, implement the Weightless Neural Network (WNN) method on a robot to identify the eyes and mouth on a static face, and enable facial recognition by capturing only the eye and mouth regions of the person being identified. The Weightless Neural Network Face Recognition Algorithm is a derivative of the WiSARD Algorithm, which focuses solely on processing facial data binary data grouped into multiple N-tuple memory units. In practice, this approach is often found in tools such as smartphones or PCs. It is also frequently implemented in robotic systems using a Raspberry Pi as the control center, supported by a microcontroller such as the Arduino Mega. The test results of this study show good performance in recognizing facial expressions, with an average accuracy of 87.96%. The expression with the highest recognition rate was “wide-open mouth” (97.34%), while the lowest accuracy was recorded for the “smiling mouth” expression (72.56%).
Inventory Code | Barcode | Call Number | Location | Status |
---|---|---|---|---|
2507005343 | T182401 | T1824012025 | Central Library (Reference) | Available but not for loan - Not for Loan |