A Hybrid Posture Detection Framework: Integrating Machine Learning and Deep Neural Networks
- Version
- Download 6
- File Size 516.84 KB
- File Count 1
- Create Date 2 February 2026
- Last Updated 2 February 2026
A Hybrid Posture Detection Framework: Integrating Machine Learning and Deep Neural Networks
Mr. B. Narasimha Rao1, N. S. Prathap2, N. Jahnavi3, N. S. S. Revanya4, U. Akash5
1Associate Professor, Dept. Of Computer Science and Engineering [2-5] B. Tech Student, Dept. Of Computer
Science and Engineering [1-5] Bonam Venkata Chalamayya Engineering College, Odalarevu
Abstract
In recent years, posture detection has become increasingly important in various fields such as healthcare, ergonomics, and human-computer interaction. This project proposes a Hybrid Posture Detection Framework that leverages the strengths of both traditional Machine Learning (ML) techniques and Deep Neural Networks (DNNs) to achieve high accuracy and robustness in identifying and classifying human postures. The framework utilizes pre-processed image or sensor data to extract relevant features, which are then analyzed through a dual-layered architecture. The first layer employs ML algorithms for preliminary posture classification based on engineered features, while th second layer refines predictions using convolutional neural networks (CNNs) to learn spatial hierarchies and complex patterns. Keywords Posture Discovery, Machine literacy, Deep Learning, Hybrid Approach, SVM, Logistic Regression, Random Forest, Naive Bayes, LSTM, 1D- CNN, 2D- CNN.
Download