Attention-based Human Activity Recognition with 3-axis Accelerometer Data Conversion

3축 가속도 데이터 변환 및 Attention 기반 사람 행동인식

A pytorch code about ETRI2023,
Attention-based Human Activity Recognition with 3-axis Accelerometer Data Conversion.

To train or inference our models, please clone this repository.😀

This project was researched by Minseong Kweon, Jaehyeong Park, Kyunghyun Kim, Jeonghyun Noh

Feel free to contact us if you have any questions,
📬 [email protected]
📬 [email protected]
📬 [email protected]
📬 [email protected]

---

Data Preprocessing

• Convert_RP.py converts time series datas of 3-axis accelerometer to RP.
  
• MFCC_convert.ipynb converts time series datas of 3-axis accelerometer to MFCC.
  

---

Training Model

$ ./run_mfcc.sh
$ ./run_rp.sh

• train_mfcc.py trains mfcc images
  
• train_rp trains rp images
  

---

Download pretrained models

Download pth files here

---

Inference

$ ./infer.sh

• infer.py tests our model (inference)
  

---

Paper Reference

[1] Ranasinghe, S., AI Machot, F., Mayr, H. C, “A review on applications of activity recognition systems with regard to performance and evaluation”, Internal Journal of Distributed Sensor Network, vol. 12 no. 8, 2016.

[2] Jaeyoung Chang, et al, “Development of Real-time Video Surveillance System Using the Intelligent Behavior Recognition Technique”, The Journal of The Institute of Internet, Broadcasting and Communication, vol. 19, no. 2, pp. 161-168, 2020.

[3] Nedorubova, A., Kadyrova, A., Khlyupin, A., “Human activity recognition using continuous wavelet transform and convolutional neural network”, doi: https://doi.org/10.48550/arXiv.2106.12666, 2021.

[4] Chen, Y., Xue, Y., “A deep learning approach to human activity recognition based on single accelerometer”, In 2015 IEEE international conference onsystems, man, and cybernetics, pp. 1488-1492, 2015.

[5] He, Z., He, Z., “Accelerometer-based Gesture Recognition Using MFCC and HMM”, In 2018 IEEE 4th International Conference on Computer and Communications (ICCC), pp. 1435-1439, 2018.

[6] Seungeun Chung, et al., “Real-world multimodallifelog dataset for human behavior study”, ETRI Journal, vol. 43, no. 6, 2021. [7] Jianjie, L., Kai-Yu, Tong, “Robust Single Accelerometer-Based Activity Recognition Using Modified Recurrence Plot”, IEEE Sensors Journal, vol. 19, no. 15, 2019.