<![CDATA[Security concerns are of utmost importance in our daily lives. Conventional door locking systems that rely on physical keys possess vulnerabilities in terms of security. Physical keys are susceptible to tampering, theft, and effortless replication. Hence, it is imperative to devise a novel approach that may effectively mitigate this issue. An example of technological use for alternative locks involves utilizing face recognition techniques to grant or deny access to doors depending on the data associated with the individual seeking entry. The primary objective of this study is to create a facial identification approach by employing machine learning techniques, namely the histogram of oriented gradients (HOG) method in conjunction with a linear Support Vector Machine (SVM). This technique is designed to be easily implemented on a Raspberry Pi 4-based Single Board Computer (SBC) that features a video sensor for machine learning input and a doorlock solenoid output. Initially, it is important to train the machine learning algorithm to accurately identify and distinguish the individual who is granted access to the door. The facial data is obtained through the capture of photographs that encompass variations in facial expression, positioning, and lighting conditions. The facial data photos are further analyzed using machine learning techniques to generate a dataset algorithm model capable of accurately identifying faces. When the system is operational and identifies a face that closely matches the trained model, the Raspberry Pi will activate the doorlock solenoid to unlock the door, and conversely, to lock the door. This approach offers security benefits as it restricts access to only those individuals whose facial features are registered in the dataset, hence allowing them to unlock the door. The developed face detection system has an accuracy rate of 83% and is compatible with computing devices possessing constrained computational capabilities, such as the SBC Raspberry Pi 4.]]>
