README.md

CPSC2020

Searching for Premature Ventricular Contraction and Supraventricular Premature Beat from Long-term ECGs: The 3rd China Physiological Signal Challenge 2020

Update on 2020/12/31

Click to view!

Tests are done on a 60s segment (median-filtered and bandpassed, sample1_fs250.mat in this folder) of a subject with frequent PVC.

• r peak detections are done using this function.
• PVC beats are labeled using red vertical lines.
• missed PVC beats are labeled using yellow boxes.
• the first image is the result by a modified version of machine learning algorithms from this repo using rr features and wavelet features, with post-processing using clinical rules. Note that phase_one_legacy is one such modified version, which uses XGBoost instead of SVM and without clinical post-processing.
• the second image is the result of the sequence labeling deep learning model with probability threshold 0.3, and filtered by a deep learning classifier. The missed PVC beats are caused by this classifier.
• the last image is the result of the sequence labeling deep learning model with probability threshold 0.5, and filtered by a deep learning classifier.

Observations

• an effective and robust rpeak (qrs complex) detector is crucial.
• the sequence labeling deep learning model (trained only for a dozen epochs because of the approaching deadline) tends to make false positive predictions but seldom has false negatives; while the deep learning classifier (trained only for several hundred epochs) has few false positives but has slightly higher probability to have false negatives.
• given a good rpeak detector, machine learning models might well be competitive against deep learning models.
• changing the threshold of the sequence labeling deep learning model from 0.5 to 0.3 can largely reduce the PVCerr score (punishment); further removing the post-filtering of the deep learning classifier might further reduce the scores, raising more false positives while reducing false negatives, considering that false negative has punishment 5 times as the punishment of false positives.

Challenge Results

Click to view!

Code No.	Institution/Affiliation	Team Members	Score PVC	Score SPB
CPSC1077	Shinall Technology	Min Chen, Kui Dong	41479	92947
CPSC1091	University of Shanghai for Science and Technology	Wenjie Cai, Jingying Yang, Jianjian Cao, Xuan Wang	55706	120942
CPSC1093	1. Soochow University; 2. Suzhou Institute of Biomedical Engineering and Technology Chinese Academy of Sciences	Lirong Wang1, Lishen Qiu2, Wenqiang Cai1, Wenliang Zhu2, Jie Yu1, Wanyue Li1, Duoduo Wang1, Huimin Zhang1	95900	111523
CPSC1082	Beijing University of Technology	Minggang Shao, Zhuhuang Zhou, Shuicai Wu	97913	95348
CPSC1089	Chengdu Spaceon Electronics CO., LTD.	Shan Yang, Chunli Wang, Heng Xiang, Qingda Kong	142228	117942
CPSC1104	Tsinghua University	Hao Wen	143484	99824
CPSC1085	Taiwan AI Academy; Academia Sinica; National Taiwan University	Tsai-Min Chen	144966	153040
CPSC1098	Northeastern University	Yan Li, Yuxiang Li, Haixu Yang, Jihong Liu	151735	215664
CPSC1092	Harbin Institute of Technology	Yang Liu, Runnan He	166215	160474
CPSC1081	Institute of Semiconductors, Chinese Academy of Sciences; University of Chinese Academy of Sciences	Yibo Yin; Sitao Zhang	168578	195467
CPSC1088	East China Jiaotong University	Feng Mei, Qian Hu, Lingfeng Liu	362348	120410

See the official website for more details.

References

1. ecg-classification
2. BioSPPy
3. Cai, Wenjie, and Danqin Hu. "QRS complex detection using novel deep learning neural networks." IEEE Access (2020).
4. torch_ecg

TODO

1. more robust qrs detector (finished)
2. feature engineering (deprecated)
3. deep learning model structure design (ongoing)
4. use SNR to deal with (eliminate?) too noisy segments?
5. etc....

NOTE

1. further updates will be done in torch_ecg, instead of this repository
2. if you find this function useful, please cite Reference 3.

Citation

DOI: 10.1088/1361-6579/ac9451