cs.LGNov 12, 2026
SaveGNN-Enhanced TCN Algorithms for ECG Anomaly Detection
Abstract
To a time series can be also given the structure of directed graph, opening the doors to the usage of graph neural networks (GNNs) in this context. In this paper we develop a new pipeline involving novel geometric deep learning models that merge the capabilities of ordinary TCNs with the ones of GNNs to solve an anomaly detection problem on electrocardiogram signals.
Figures & tables
Figure 1: One possible structure of a time-digraph, as described in Example 1. Here only adjacent connections and all the connections for node are shown, and
Figure 2: The encoder and decoder (a), and the skip-connections block (b)
Figure 3: The structure of the autoencoder.
Figure 4: Representations of the hidden embeddings of the signals obtained by TGraphMixedVAE_hdcl (a) and TCNGraphVAE1_hdcl (b), and converted to 2D representations using TSNE. Label indicates a normal signal, an abnormal signal.
More in this topic
Time-Series Anomaly Detection
Time-Series Anomaly Detection identifies unusual patterns or events in sequential data, from sensor readings and network traffic to health and industrial telemetry.
See all 226 papers