Rui Pedro Carrilho

Mr. Rui Pedro Carrilho

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 September 2024 in Electronics
Verified
Journal Article
Open access
PDF file
RCRui Pedro Carrilho
Rui Pedro Carrilho
EYEhsan Yaghoubi
Ehsan Yaghoubi
JLJosé Lindo
José Lindo
KHKailash A. Hambarde
Kailash A. Hambarde
Hugo ProençaHugo Proença
Hugo Proença
Defect detection is a crucial part of the pipeline in many industries. In the textile industry, it is especially important, as it will affect the quality and price of the final product. However, it is mostly performed by human agents, who have been reported to have poor performance, along with requiring a costly and time-consuming training process. As such, methods to automate the process have been increasingly explored throughout the last 20 years. While there are many traditional approaches to this problem, with the advent of deep learning, machine learning-based approaches now constitute the majority of all possible approaches. Other articles have explored traditional approaches and machine learning approaches in a more general way, detailing their evolution over time. In this review, we summarize the most important advancements in the last 5 years and focus mostly on machine learning-based approaches. We also outline the most promising avenues of research in the future.

9 April 2025 in Sensors
Verified
Journal Article
Open access
PDF file
ALASMAE LAMSAF
ASMAE LAMSAF
RCRui Pedro Carrilho
Rui Pedro Carrilho
João C. NevesJoão C. Neves
João C. Neves
Hugo ProençaHugo Proença
Hugo Proença
Causality, which involves distinguishing between cause and effect, is essential for understanding complex relationships in data. This paper provides a review of causality in two key areas: causal discovery and causal inference. Causal discovery transforms data into graphical structures that illustrate how variables influence one another, while causal inference quantifies the impact of these variables on a target outcome. The models are more robust and accurate with the integration of causal reasoning into machine learning, improving applications like prediction and classification. We present various methods used in detecting causal relationships and how these can be applied in selecting or extracting relevant features, particularly from sensor datasets. When causality is used in feature selection, it supports applications like fault detection, anomaly detection, and predictive maintenance applications critical to the maintenance of complex systems. Traditional correlation-based methods of feature selection often overlook significant causal links, leading to incomplete insights. Our research highlights how integrating causality can be integrated and lead to stronger, deeper feature selection and ultimately enable better decision making in machine learning tasks.

19 June 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
RCRui Pedro Carrilho
Rui Pedro Carrilho
KHKailash A. Hambarde
Kailash A. Hambarde
Hugo ProençaHugo Proença
Hugo Proença
Detecting anomalies in texture has become a significant concern across various industrial processes. One prevalent application of this is in inspecting patterned textures, especially in the domain of fabric defect detection, which is a commonly encountered scenario. This task entails dealing with a wide array of colours and textile varieties, spanning a broad spectrum of fabrics. Due to the extensive diversity in colours, textures, and defect characteristics, fabric defect detection presents a complex and formidable challenge within the realm of patterned texture inspection. While recent trends have seen a rise in the utilization of deep learning methods for anomaly detection, there still exist notable gaps in this field. In this paper, we introduce a novel dataset comprising a diverse selection of fabrics and defects from a textile company based in Portugal. Our contributions encompass the provision of this unique dataset and the evaluation of state-of-the-art (SOTA) methods’ performance on our dataset.

24 September 2026 in Electronics
Verified
Journal Article
Open access
PDF file
RCRui Pedro Carrilho
Rui Pedro Carrilho
MD RashidunnabiMD Rashidunnabi
MD Rashidunnabi
Hugo ProençaHugo Proença
Hugo Proença
Reliable fabric inspection is difficult to automate because production-line imagery differs substantially from the controlled samples used by many public benchmarks. This paper introduces LUSITANOv2, a high-resolution fabric-defect dataset collected at an active textile inspection station with an industrial line-scan camera and directional illumination. It contains 25,120 native images, including both defect-containing and defect-free fabric, together with 18,557 class-agnostic bounding boxes. We establish supervised baselines with YOLOv12n, Faster R-CNN, and RT-DETR-L and evaluate nine one-class anomaly-detection methods. RT-DETR-L achieves the highest mAP@0.5 in the matched detector comparison, reaching 0.610, compared with 0.597 for Faster R-CNN and 0.592 for YOLOv12n. In a separate controlled YOLOv12n resolution study, retaining more native weave detail improves detection performance, reaching 0.598 mAP@0.5 at the highest tested resolution. A post hoc analysis finds no simple recall penalty for defects that touch the image boundary. Matched zero-shot experiments across LUSITANOv2, TILDA, and ZJU-Leaper reveal substantial cross-dataset degradation despite stronger in-domain results. LUSITANOv2 provides a realistic benchmark for localized defect detection, one-class anomaly detection, resolution-sensitive line-scan processing, and future work on transfer to active textile-production settings.