Related Experiment Video
Updated: Sep 18, 2025
07:47
Design and Evaluation of Smart Glasses for Food Intake and Physical Activity Classification
Published on: February 14, 2018
11.4K
UCA-YOLOv8n: a real-time and efficient fruit chunks detection algorithm for meal-assistance robot.
1Shanghai Polytechnic University, Shanghai, China.
Peerj. Computer Science
|June 26, 2025
Summary
This study introduces an enhanced YOLOv8n algorithm for accurate fruit chunk detection in assistive robotics. The improved UCA-YOLOv8n model offers real-time performance with reduced size and increased detection accuracy.
Area of Science:
- Computer Vision
- Robotics
- Artificial Intelligence
Background:
- Assistive technologies require efficient object detection for tasks like meal assistance.
- Current fruit chunk detection algorithms lack speed and precision for robotic applications.
Purpose of the Study:
- To develop an optimized YOLOv8n algorithm for real-time, high-accuracy fruit chunk detection.
- To enhance object detection capabilities for meal-assistance robots.
Main Methods:
- Integrated Universal Inverted Bottleneck (UIB) module to reduce parameters.
- Incorporated coordinate attention (CA) mechanism for improved focus on target regions.
- Embedded ADown module from YOLOv9 into the YOLOv8 backbone network.
Main Results:
- Achieved a 1.9 MB size reduction and 2.5 GFLOPs decrease.
- Increased mAP50 by 2.1% and mAP50-95 by 3.3%.
- Demonstrated superior performance compared to mainstream object detection algorithms.
Conclusions:
- The UCA-YOLOv8n algorithm enables real-time, accurate detection of fruit chunks.
- The proposed enhancements significantly improve detection accuracy and reduce model size.
- This advancement is crucial for developing more capable meal-assistance robots.
