Abstract:The primary objective of this research investigation is to examine object detection within the specific environment of a nursery. The nursery environment presents a complex scene with a multitude of objects, varying in size and background. To simulate real-world conditions, we gathered data from a nursery. Our study is centered around the detection of small objects, particularly in nursery settings where objects that include stationery, toys, and small accessories are commonly present. These objects are of significant importance in facilitating cognition of the activities and interactions taking place within the confines of the room. Due to their small size and the possibility of occlusion by other objects or children, precisely detecting these objects is regrettably fraught with inherent challenges. This study introduces YOLOv8s-NE in an effort to enhance the detection of small objects found in the nursery. We improve the standard YOLOv8 by incorporating an extra detection head to effectively for small objects. We replace the C2f module with C2f_DCN to further improve the model's ability to detect objects of varying sizes that can be deformed or occluded within the image. Furthermore, we introduce NAM attention to focus on the important features and ignore less informative ones, thereby improving the accuracy of our proposed model. We used the five-fold cross-validation approach to split the dataset in order to evaluate the performance of YOLOv8s-NE, thereby facilitating a more comprehensive model evaluation. Our model achieves 34.1% of APs, 45.1% of mAP50:90, and 76.7% of mAP50 detection accuracy at 37.55 FPS on the nursery dataset. In terms of APs, mAP50:90, and mAP50 metrics, our proposed YOLOv8s-NE model outperforms the standard YOLOv8s model, with improvements of 4.6%, 4.7%, and 3.9%, respectively. We apply our proposed YOLOv8s-NE model as a safety system by developing an algorithm to detect objects on top of cabinets that could be potentially risky to children.

Design of A Real-Time Object Detection Prototype System with YOLOv3 (You Only Look Once)

Image Detection and Real Time Object Detection

Real Time Object Detection System Using Yolov

Design and implementation of real-time object detection system based on single-shoot detector and OpenCV

ObjectDetect: A Real-Time Object Detection Framework for Advanced Driver Assistant Systems Using YOLOv5

Design of Real-Time Object Detection in Mobile Robot for Volcano Monitoring Application

A Learning Framework for Target Detection and Human Face Recognition in Real Time

MobileYOLO: Real-Time Object Detection Algorithm in Autonomous Driving Scenarios

Accelerating Object Detection with YOLOv4 for Real-Time Applications

Real-Time Object Detection in Occluded Environment with Background Cluttering Effects Using Deep Learning

Statistical Analysis of Design Aspects of Various YOLO-Based Deep Learning Models for Object Detection

You Only Look Once: Unified, Real-Time Object Detection

Real-time object detection and segmentation technology: an analysis of the YOLO algorithm

YOLOv8: A Novel Object Detection Algorithm with Enhanced Performance and Robustness

Object detection using YOLO: challenges, architectural successors, datasets and applications

A Real-Time Object Detector for Autonomous Vehicles Based on YOLOv4

PLOD-YOLO: Premium Lightweight Object Detection for Autonomous Following Robot

YOLO9000: Better, Faster, Stronger

An Object Detection System Based on YOLO in Traffic Scene

I-YOLO: a novel single-stage framework for small object detection

YOLOv8s-NE: Enhancing Object Detection of Small Objects in Nursery Environments Based on Improved YOLOv8