Go Beyond Multiple Instance Neural Networks: Deep-learning Models based on Local Pattern Aggregation

TL;DR

This paper introduces LPANet, a novel deep-learning model that combines cropping and aggregation operations to handle variable-sized data efficiently, improving generalization over traditional CNNs and LSTMs.

Contribution

The paper proposes LPANet, a new network architecture with local pattern aggregation for variable-size data, enhancing performance and ease of tuning compared to existing models.

Findings

LPANet outperforms classical CNN and LSTM models in ventricular contraction detection.

LPANet reduces parameter tuning difficulty and improves generalization.

Experimental results demonstrate the effectiveness of LPANet on ECG data.

Abstract

Deep convolutional neural networks (CNNs) have brought breakthroughs in processing clinical electrocardiograms (ECGs), speaker-independent speech and complex images. However, typical CNNs require a fixed input size while it is common to process variable-size data in practical use. Recurrent networks such as long short-term memory (LSTM) are capable of eliminating the restriction, but suffer from high computational complexity. In this paper, we propose local pattern aggregation-based deep-learning models to effectively deal with both problems. The novel network structure, called LPANet, has cropping and aggregation operations embedded into it. With these new features, LPANet can reduce the difficulty of tuning model parameters and thus tend to improve generalization performance. To demonstrate the effectiveness, we applied it to the problem of premature ventricular contraction detection and the experimental results shows that our proposed method has certain advantages compared to classical network models, such as CNN and LSTM.