Enhancing Frame-Level Stability in YOLO Image Classification with Simple Moving Average for Hand Gesture Recognition
- Also see (Korean ver.): Google Slide Presentation (한국어 발표 자료)
This repository contains the implementation of a hand gesture classification system using YOLO v8 Nano. The project aims to enhance the frame-level stability in image classification by applying a simple moving average (SMA) technique to smooth out oscillations in the classification of hand gestures.
- Object Classification: Utilizes YOLO v8 Nano for efficient and fast hand gesture recognition.
- Oscillation Reduction: Implements SMA to reduce frame-to-frame classification fluctuations.
- Three-Class System: Classifies hand gestures into 'open hand', 'folding hand', and 'folded hand'.
-
Frame-Level Oscillation Reduction: Employs a simple moving average (SMA) strategy to mitigate frame-to-frame oscillations, leading to more stable and consistent gesture recognition.
-
Enhanced Accuracy and Stability: The SMA technique significantly improves the accuracy and stability of gesture classification, especially in scenarios where traditional methods might struggle with rapid class fluctuations.
-
Interactive Virtual/Augmented Reality: Can be integrated into VR/AR systems for more natural and intuitive hand gesture controls, enhancing user experience in virtual environments.
-
Smart Home Control: Suitable for gesture-based control systems in smart homes, allowing users to manage devices or settings with simple hand movements.
-
Accessible Computer Interfaces: Offers an alternative way for individuals with disabilities to interact with computers and digital devices using hand gestures.
-
Educational Tools: Can be used in educational software, especially in interactive learning environments, to make educational content more engaging and interactive.
-
Gesture-Controlled Robotics: Applicable in controlling robots or automated systems where hand gestures can be used for commands, making the interaction more intuitive and user-friendly.
These highlights and use cases illustrate the versatility and practicality of your project, emphasizing its potential impact in various technology sectors.
- Python 3.9
- Dependencies listed in
requirements.txt
- Clone the repository:
git clone https://github.com/dchlseo/yolo-moving-average.git - Install dependencies:
pip install -r requirements.txt
To run the gesture classification:
(TO BE UPDATED)
This README provides a concise yet comprehensive guide to setting up and using the project. Modify it as necessary to fit the specific details and requirements of your implementation.