A Comparative Research on Human Activity Recognition Using Deep Learning
| Autor: | Ozen Ozkaya, Nilay Tufek |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Computer science
business.industry Deep learning 010401 analytical chemistry 02 engineering and technology Machine learning computer.software_genre 01 natural sciences Convolutional neural network 0104 chemical sciences Activity recognition Comparative research 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence Layer (object-oriented design) business computer |
| Zdroj: | SIU |
| DOI: | 10.1109/siu.2019.8806395 |
| Popis: | In recent years, action recognition is becoming more popular in many fields such as person surveillance, human-robot interaction due to the widespread usage of various sensors. In this study, we aimed to develop an action recognition system that is intended to recognize human actions by using only accelerometer and gyroscope data. Various deep learning approaches like Convolutional Neural Network(CNN), Long-Short Term Memory (LSTM) with classical machine learning algorithms and their combinations were implemented and evaluated. A data augmentation method were applied while accuracy rates were increased noticeably.%98 accuracy rate obtained by using 3 layer LSTM network which means a solid contribution. Additionally, a realtime application was developed by using LSTM network. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|