A simplified network topology for fruit detection, counting and mobile-phone deployment

Lawal, Olarewaju Mubashiru; Zhu, Shengyan; Cheng, Kui; Liu, Chuanli

doi:10.1371/journal.pone.0292600

Cited by 2 publications

(2 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Dalam penelitian ini, digunakan format NCNN yang dikembangkan oleh Tencent karena performanya yang unggul dalam komputasi neural network (Lawal et al, 2023). NCNN telah dikembangkan dan diunggah ke ncnn-android-yolov8 (FeiGeChuanShu & Q-Engineering, 2023), dan digunakan pada perangkat Huawei Nova 10 Pro.…”

Section: Integrasi Modelunclassified

Pengembangan Sistem Deteksi On-Shelf Availability Produk Menggunakan Algoritma YOLOV8 pada Aplikasi Beregerak

Andaru,

Fudholi

2024

jimik

View full text Add to dashboard Cite

Research related to retail operations has been a major focus in recent years, driven by rapidly changing market dynamics and the importance of product availability on store shelves to meet customer satisfaction. The concept of On Shelf Availability (OSA) has become key in ensuring products are available when needed. However, the challenge in retail management lies in monitoring thousands of different products, which is time-consuming and resource-intensive. To address this issue, an efficient object detection solution is needed. The research implements the YOLOv8 algorithm in detecting out-of-stock items, particularly in the context of mobile devices with resource limitations. In order to achieve this goal, the research adopts a comprehensive methodology, starting from direct data collection from supermarkets, data processing, labeling, to model training using transfer learning techniques. Transfer learning method is chosen to overcome data limitations and accelerate the model training process, enabling faster adaptation to object detection conditions at specific locations. Test results show that YOLOv8s delivers the best performance with an accuracy of up to 94.7%, allowing real-time object detection. Testing is conducted on various mobile devices, including Samsung A54 and Samsung A6, where YOLOv8n consistently performs with an inference time of 41.46 ms on Samsung A54 and 257.73 ms on Samsung A6. The main contribution of this research is to enhance object detection capabilities on devices with low computational power, such as mobile devices, and provide an effective solution to the problem of product availability on store shelves. Thus, this research not only brings positive impact to retail management but also drives the development of object detection technology in the context of resource-limited devices and unstable internet connections.

show abstract

Section: Integrasi Modelunclassified

Pengembangan Sistem Deteksi On-Shelf Availability Produk Menggunakan Algoritma YOLOV8 pada Aplikasi Beregerak

Andaru,

Fudholi

2024

jimik

View full text Add to dashboard Cite

show abstract

“…Nevertheless, the weight-size of the modified YOLOv5 algorithm is still large, and the detection accuracy needs further improvement. Lawal et al [28] proposed a simplified network that could detect fruit targets at 82.4% AP and an average speed of 461 FPS, but had a large parameter. Similarly, the real-time cucurbit fruit detection in greenhouses using an improved YOLOv5 by Lawal [29] and YOLO-Banana proposed by Fu et al [30] recorded excellent performance but had a large weight-size.…”

Section: Introductionmentioning

confidence: 99%

Lightweight fruit detection algorithms for low‐power computing devices

Lawal,

Zhao,

Zhu

et al. 2024

IET Image Processing

Self Cite

View full text Add to dashboard Cite

A lightweight fruit detection algorithm is important to ensure real‐time detection on low‐power computing devices while maintaining detection accuracy. In addition, the fruit detection algorithm is also faced with some environmental factors. To solve these challenges, lightweight detection algorithms termed YOLO‐Lite, YOLO‐Liter and YOLO‐Litest were developed based on the YOLOv5 framework. The compared mean average precision (mAP) detection revealed that YOLO‐Lite at 0.86 is 2%, 4%, 5%, 7%, and 16% more than YOLO‐Liter and YOLOv5n at 0.84 each, YOLOv4‐tiny at 0.82, YOLO‐Liter at 0.81, YOLO‐MobileNet at 0.79, and YOLO‐ShuffleNet at 0.70, respectively, but not for YOLOv8n at 0.87. On the Computer platform, except for YOLOv4‐tiny at 178.6 frames per second (FPS), the speed of YOLO‐Litest at 158.7 FPS is faster than YOLO‐Liter at 129.9 FPS, YOLO‐Lite at 120.5 FPS, YOLO‐ShuffleNet at 119.0 FPS, YOLOv8n at 116 FPS, YOLOv5n at 111.1 FPS, and YOLO‐MobileNet at 89.3 FPS. Using Jetson Nano, the 32.3 FPS of YOLO‐Litest is faster than other algorithms, but not YOLOv4‐tiny's 34.1 FPS. On the Raspberry Pi 4B, YOLO‐Litest with 4.69 FPS, outperformed other algorithms. The choices for an accurate and faster detection algorithm are YOLO‐Lite and YOLO‐Litest respectively, while YOLO‐Liter maintains a balance between them.

show abstract

A simplified network topology for fruit detection, counting and mobile-phone deployment

Cited by 2 publications

References 41 publications

Pengembangan Sistem Deteksi On-Shelf Availability Produk Menggunakan Algoritma YOLOV8 pada Aplikasi Beregerak

Pengembangan Sistem Deteksi On-Shelf Availability Produk Menggunakan Algoritma YOLOV8 pada Aplikasi Beregerak

Lightweight fruit detection algorithms for low‐power computing devices

Contact Info

Product

Resources

About