Browsing by Author "Arce Castro, Josue David"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Identificación de defectos en textiles utilizando algoritmos de inteligencia artificial
    (Universidad de Cuenca, 2025-03-20) Arce Castro, Josue David; Llivisaca Villazhañay, Juan Carlos
    Timely and accurate detection of defects in textile products is crucial to ensure quality, reduce costs and meet consumer demands. However, traditional visual inspection methods have limitations in terms of consistency, scalability and objectivity. This research evaluated the potential of several unsupervised anomaly detection algorithms based on artificial intelligence techniques to improve textile defect identification. The performances of Isolation Forest, Local Outlier Factor (LOF), Elliptic Envelope, One-Class SVM, DBSCAN and K-Means were implemented and compared. The results showed that the Isolation Forest model stood out for obtaining the best values in accuracy, precision and F1-score, indicating its ability to detect defects with few false positives and negatives. On the other hand, LOF had the highest sensitivity, while DBSCAN presented a relatively high specificity. The implementation of these AI algorithms in textile production environments could automate and optimize inspection processes, reducing costs and improving product quality. This research contributes to the field of artificial intelligence applied to manufacturing, demonstrating the potential of anomaly detection models to address specific challenges in the textile industry.