Objective . Although deep learning-based current methods have achieved impressive results in electrocardiograph (ECG) arrhythmia classification issues, they rely on using the original data to identify arrhythmia categories. However, a large amount of data generated by long-term ECG monitoring pose a significant challenge to the limited-bandwidth and real-time systems, which limits the application of deep learning in ECG monitoring. Approach . This paper, therefore, proposed a novel multi-task network that combined compressed sensing and convolutional neural networks, namely CSML-Net. According to the proposed model, the ECG signals were compressed by utilizing a learning measurement matrix and then recovered and classified simultaneously via shared layers and two task branches. Among them, the multi-scale feature module was designed to improve model performance. Main results . Experimental results on the MIT-BIH arrhythmia dataset demonstrate that our proposed method is superior to all the approaches that have been compared in terms of reconstruction quality and classification performance. Significance . Consequently, the proposed model achieving the reconstruction and classification in the compressed domain can be an improvement and become a promising approach for ECG arrhythmia reconstruction and classification.

Deep learning (DL) models have made remarkable progress in various sequence processing tasks. It is widely acknowledged that these models heavily rely on numerous training data and finely tuned parameters. Recent studies highlighted that conventional convolutions in deep networks may hamper feature processing efficacy, particularly in long temporal sequence analysis, due to their limited feature representation capabilities. To tackle these challenges, this article introduces a novel generalizable convolutional neural network (GeCNN) architecture tailored for temporal domain long sequence identification. Our framework incorporates three key components: the generic convolutional neural network (CNN), selective CNN, and multiple pooling layers. The generic CNN implements customizable hyper-convolutional operations through non-linear convolvers, thereby enhancing feature representation effectiveness and significantly improving accuracy. Subsequently, the selective CNN is designed to abate the demand for large training data by focusing on various subsequences. We propose the homogeneous striding principle and the partial homogeneous striding theorem to theoretically support the method. The multiple pooling combines eight distinct pooling operations to mitigate the statistical information loss problem typically associated with single pooling actions. Experimental results demonstrate that our GeCNN architecture achieves superior performance with shallow networks and tiny data compared to existing deep networks. The accuracy of the best-trained model surpasses the ResNet and self-attention-based models by $9.51\%$ and $16.79\%$ utilizing only $0.18\%$ of data for training in the GTZAN dataset. Meanwhile, the accuracy of the optimal model overtakes the other two models by $5.35\%$ and $10.16\%$ while using merely $1.56\%$ of data for training in the PLAID dataset.