Security System for Smart Homes to Prevent Theft

2. Related Works

The absence of smart devices and sophisticated facial recognition software at home is contributing to an increase in theft tracking (Ahmed et.al. 2010). The inaccuracy of face detection techniques acquired by CCTV cameras arises from theft victims hiding their faces, either completely or partially, using materials made of leather or cloth. Numerous researchers have directed their efforts towards improving face recognition technology (Jian et.al.2018). For example, Zhiwei et al. (Zhang et.al.2012) introduced a regularization technique aimed at enhancing face identification across different spectrums. Additionally, many researchers have focused on refining face and eye detection using the Haar classifier algorithm. Xin et al. (Xin et.al 2018) proposed a system for person face recognition, claiming it offers superior security compared to unimodal biometric systems. Nguyen et al. Nguyen et.al. 2018 suggested employing deep learning techniques, while Cho et al. Cho et.al 2018 advocated for the use of convolutional neural networks to improve image features in face recognition systems, particularly those equipped with visible light camera sensors. Binary pattern techniques were used in the creation of the face detection system by Alobaidi et al.2018. Zhang et al.'s approach to face detection in a stable environment combines model-driven and data-driven methods. Using support vector machines, Omid et al. created a face recognition system that works in the presence of cosmetics like contact lenses and facial makeup (Sharifi et.al.2018). Door and window security is a major consideration when it comes to house security. These days, digital doors that don't require a physical key are employed because to advancements in IoT technology. Nevertheless, it is simple to damage digital doors, and the owners are only aware of theft once they get home. Huth et al. devised a wirelessly connected security system employing a physical key generation method, demonstrating its application in smart homes (Andreasa et.al.2019). To enhance energy efficiency, WiFi modules, temperature sensors, and door sensors are commonly employed.

Intruder detection systems often utilize laser and LDR sensors to detect movement, with alerts sent to homeowners via SMS (Cristian et.al.2016). However, this method may fail to transmit messages through a GSM module in areas lacking internet coverage. Anitha et al. proposed an artificial intelligence-based home system (Anitha 2016).

Patel et al. introduced a modern door lock system for homes, incorporating a Raspberry Pi system-on-chip (SoC), a camera, and an infrared sensor (Jay Patel et.al.2019). This system operates by granting access to individuals whose images are stored in the cloud. Nivo Suranth et al. implemented a home security system utilizing a PIR sensor, an Arduino microcontroller, and a Raspberry Pi 3 SoC (Nicosurantha 2018). The Raspberry Pi processes images captured by a webcam connected to the SoC, along with sensor data from the Arduino microcontroller. Additionally, intruders can be identified using a support vector algorithm, capable of detecting intruders within 2 seconds.