Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhanced Self-Checkout System for Retail Based on Improved YOLOv10

Lianghao Tan, Shubing Liu, Jing Gao, Xiaoyi Liu, Linyue Chu, Huangqi Jiang

TL;DR

The paper addresses the need for fast, accurate product recognition in self-checkout by introducing MidState-YOLO-ED, an enhanced detector that fuses a YOLOv8-style head with a YOLOv10 backbone and adds EMA attention and lightweight C2f-Dual design. It demonstrates improved mAP and recall on a retail-like RPC dataset, achieving a compact, real-time model suitable for embedded hardware. The approach is validated through ablation studies and comparisons against Faster R-CNN and SSD, showing a favorable speed-accuracy trade-off and a substantial reduction in parameters and FLOPs. The work offers practical impact for self-checkout efficiency, inventory control, and customer service in retail, with potential deployment on resource-constrained devices.

Abstract

With the rapid advancement of deep learning technologies, computer vision has shown immense potential in retail automation. This paper presents a novel self-checkout system for retail based on an improved YOLOv10 network, aimed at enhancing checkout efficiency and reducing labor costs. We propose targeted optimizations to the YOLOv10 model, by incorporating the detection head structure from YOLOv8, which significantly improves product recognition accuracy. Additionally, we develop a post-processing algorithm tailored for self-checkout scenarios, to further enhance the application of system. Experimental results demonstrate that our system outperforms existing methods in both product recognition accuracy and checkout speed. This research not only provides a new technical solution for retail automation but offers valuable insights into optimizing deep learning models for real-world applications.

Enhanced Self-Checkout System for Retail Based on Improved YOLOv10

TL;DR

The paper addresses the need for fast, accurate product recognition in self-checkout by introducing MidState-YOLO-ED, an enhanced detector that fuses a YOLOv8-style head with a YOLOv10 backbone and adds EMA attention and lightweight C2f-Dual design. It demonstrates improved mAP and recall on a retail-like RPC dataset, achieving a compact, real-time model suitable for embedded hardware. The approach is validated through ablation studies and comparisons against Faster R-CNN and SSD, showing a favorable speed-accuracy trade-off and a substantial reduction in parameters and FLOPs. The work offers practical impact for self-checkout efficiency, inventory control, and customer service in retail, with potential deployment on resource-constrained devices.

Abstract

With the rapid advancement of deep learning technologies, computer vision has shown immense potential in retail automation. This paper presents a novel self-checkout system for retail based on an improved YOLOv10 network, aimed at enhancing checkout efficiency and reducing labor costs. We propose targeted optimizations to the YOLOv10 model, by incorporating the detection head structure from YOLOv8, which significantly improves product recognition accuracy. Additionally, we develop a post-processing algorithm tailored for self-checkout scenarios, to further enhance the application of system. Experimental results demonstrate that our system outperforms existing methods in both product recognition accuracy and checkout speed. This research not only provides a new technical solution for retail automation but offers valuable insights into optimizing deep learning models for real-world applications.
Paper Structure (20 sections, 5 equations, 3 figures, 3 tables)

This paper contains 20 sections, 5 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Research Background
  3. Challenges in Current Self-Checkout Systems
  4. Overview of Advancements of Deep Learning and Computer Vision in Retail
  5. Purpose of the New System and Advantages
  6. Related Work
  7. Traditional Object Detection Algorithms
  8. Development of the YOLO Series
  9. Application of YOLO
  10. The Improved MidState-YOLO-ED Network
  11. Integration of YOLOv8 and YOLOv10
  12. Integration of EMA Attention
  13. Lightweight Dual Convolution C2f-Dual Design
  14. Experimental Results and Analysis
  15. Experimental Setup and Parameters
  16. ...and 5 more sections

Figures (3)

  • Figure 1: Dual Convolution C2f-Dual Design
  • Figure 2: Dual Convolution C2f-Dual Design
  • Figure 3: Dual Convolution C2f-Dual Design