One-lead ischemia detection using a new backpropagation algorithm and the European ST-T database

SUTIL: Intelligent ischemia monitoring system

“…The real time monitoring of patients with ischemic cardiopathies clearly adds versatility to the treatment of these cardiopathies which cannot be obtained by using other off line monitoring techniques such as, for example, Holter monitoring. Even though there are other techniques which offer a similar performance to that obtained by our system [28,29], they are not usually suitable for real time implementation, unless specific hardware is used.…”

An ischemia detection method based on artificial neural networks

“…It must be noted that the majority of the results reported in Table 4 refer to subsets of ECG recordings of the ESC ST-T database [27,[29][30][31] and only some have used all the ESC ST-T database recordings [13,23,26]. Also, the currently evaluated beat classification network performs better than similar approaches [28], as indicated in Table 1.…”

“…Fuzzy expert systems [31,32] manage to keep this feature without applying strict threshold values. ANNs [4,21,[26][27][28][29] due to their non-linear characteristics and learning capabilities have provided good performance results. The above methods when tested with the ESC ST-T database demonstrated a Se that ranged from 71 to 94% and a PPA from 66 to 90%.…”

“…Se a (%) PPA b (%) Jager et al [13] 85 86 Papaloukas et al [23] 94 79 Silipo and Marchesi [26] 77 82 Silipo and Marchesi [26] 77 83 Silipo and Marchesi [26] 77 85 Silipo and Marchesi [26] 71 66 Silipo and Marchesi [26] 77 86 Silipo et al [27] 85 88 Silipo et al [27] 78 90 Stamkopoulos et al [29] 84 79 Taddei et al [30] 84 85 Vila et al [31] 83 A disadvantage of the proposed method is that it cannot provide any interpretation, for the decisions made due to the employment of the neural network model. Proper combination [18] of the knowledge-based approach [23] with the neural network model can eliminate this drawback.…”

“…The accuracy of beat classification influences ischemic episode definition where sequences of ischemic beats need to be identified. Various methods have been proposed for ischemia detection based on set of rules [2,5,15,20,[25][26][27], artificial neural networks (ANNs) [4,21,[26][27][28][29], fuzzy logic [31,32] or other signal analysis techniques [3,11,13,16,30].…”