IoT Based Smart Mirror Using Raspberry Pi 4 and YOLO Algorithm: A Novel Framework for Interactive Display

Abstract

Objectives:The primary goal of this device is to support the user in a variety of ways, and by enabling communication between the user and the gadget, we can demonstrate the user's control over the location where the device will be Built and set up the home environment for the human detection using You Only Look Once (YOLO) algorithm. Methods: The technology is made to show the current news, weather, and temperature on the mirror. The technology is primarily intended to be used as an intrusion detection system and for human monitoring. YOLO algorithm is used to find the object in the given image. The suggested design is envisioned as a collection of components that can be utilized to provide security as well as simply display information on a screen. The system is created utilizing Python programming and hardware components such as Raspberry Pi 3 model, a microphone, a touch screen, a mobile device, a camera, and passive infrared (PIR) sensors. Findings: This study provides thorough information regarding the operation of a smart mirror, which enables us to perform all necessary actions or tasks that we perform regularly and to complete them in a predetermined amount of time. Since the research is primarily intended for intrusion detection, we can also get the most out of it by using it in a specific manner. We would use it to monitor children's whereabouts. Novelty: This technology has a high economic value because it has so many uses and tends to help users with daily tasks. And because it is created using machine learning language and has room for artificial intelligence, even the user who uses it can add some of his remainders to the device, and it can also alert a user in the designated remainder time so that person won't miss their important meeting. As a result, this device is more valuable, useful, and collectively beneficial. This technology will produce a finished product with the highest level of accuracy and precision.

An Object Detection Approach for Automated Detection of Groove Line in Tube Yoke

“…Training YOLOv8 tiny and small models on a custom dataset for 1000 epochs with early stop. YOLOv8 small stopped after 278 and tiny after 219 epochs this is due to the small data set [21]. Validating small and tiny models in both the file formats on the custom validation set, results are as follows.…”

“…We trained tiny and small architecture. This is because of the constraints on model size, computational resources, and minimum inference time required [21].…”

“…The background study of various object detection algorithms is covered in Section 2 of this article. The design and architecture of YOLOv8 are the main topics of section 3 [21]. In section 4, we go over our suggested methodology, which outlines the procedures for detecting the groove, format conversion, and Raspberry Pi deployment.…”

A Real-Time Small Target Vehicle Detection Algorithm with an Improved YOLOv5m Network Model

“…It predicts the final target box by generating a set of candidate boxes, and then classifying and regressing these boxes. The second type is the one-stage detection algorithm using regression, with YOLO as its typical representative [8][9][10][11]. It directly convolves and pools the image to generate candidate boxes, and performs classification and regression at the same time to detect the vehicle object.…”

A mobile device framework for video captioning using multimodal neural networks

“…In [13], the authors have developed a smart device utilizing ARM-based processor hardware for video surveillance. The system incorporates a YOLO network that receives data captured by a video camera triggered by motion detected through an infrared sensor.…”