Showing 1–2 of 2 results for author: Arizmendi, C J

Automatic welding detection by an intelligent tool pipe inspection

Authors: C J Arizmendi, W L Garcia, M A Quintero

Abstract: This work provide a model based on machine learning techniques in welds recognition, based on signals obtained through in-line inspection tool called smart pig in Oil and Gas pipelines . The model uses a signal noise reduction phase by means of preprocessing algorithms and attributeselection techniques. The noise reduction techniques were selected after a literature review and testing with survey… ▽ More This work provide a model based on machine learning techniques in welds recognition, based on signals obtained through in-line inspection tool called smart pig in Oil and Gas pipelines . The model uses a signal noise reduction phase by means of preprocessing algorithms and attributeselection techniques. The noise reduction techniques were selected after a literature review and testing with survey data. Subsequently, the model was trained using recognition and classification algorithms, specifically artificial neural networks and support vector machines. Finally, the trained model was validated with different data sets and the performance was measured with cross validation and ROC analysis. The results show that is possible to identify welding automatically with an efficiency between 90 and 98 percent △ Less

Submitted 11 March, 2025; originally announced March 2025.

Development of a Deep Learning Model for the Prediction of Ventilator Weaning

Authors: Hernando Gonzalez, Carlos Julio Arizmendi, Beatriz F. Giraldo

Abstract: The issue of failed weaning is a critical concern in the intensive care unit (ICU) setting. This scenario occurs when a patient experiences difficulty maintaining spontaneous breathing and ensuring a patent airway within the first 48 hours after the withdrawal of mechanical ventilation. Approximately 20 of ICU patients experience this phenomenon, which has severe repercussions on their health. It… ▽ More The issue of failed weaning is a critical concern in the intensive care unit (ICU) setting. This scenario occurs when a patient experiences difficulty maintaining spontaneous breathing and ensuring a patent airway within the first 48 hours after the withdrawal of mechanical ventilation. Approximately 20 of ICU patients experience this phenomenon, which has severe repercussions on their health. It also has a substantial impact on clinical evolution and mortality, which can increase by 25 to 50. To address this issue, we propose a medical support system that uses a convolutional neural network (CNN) to assess a patients suitability for disconnection from a mechanical ventilator after a spontaneous breathing test (SBT). During SBT, respiratory flow and electrocardiographic activity were recorded and after processed using time-frequency analysis (TFA) techniques. Two CNN architectures were evaluated in this study: one based on ResNet50, with parameters tuned using a Bayesian optimization algorithm, and another CNN designed from scratch, with its structure also adapted using a Bayesian optimization algorithm. The WEANDB database was used to train and evaluate both models. The results showed remarkable performance, with an average accuracy 98 when using CNN from scratch. This model has significant implications for the ICU because it provides a reliable tool to enhance patient care by assisting clinicians in making timely and accurate decisions regarding weaning. This can potentially reduce the adverse outcomes associated with failed weaning events. △ Less

Submitted 3 March, 2025; originally announced March 2025.