Detecting small targets and handling target occlusion and overlap are critical challenges in weld defect detection. In this paper, we propose the S-YOLO model, a novel weld defect detection method based on the YOLOv8-nano model and several mathematical techniques, specifically tailored to address these issues. Our approach includes several key contributions. Firstly, we introduce omni-dimensional dynamic convolution, which is sensitive to small targets, for improved feature extraction. Secondly, the NAM attention mechanism enhances feature representation in the region of interest. NAM computes the channel-wise and spatial-wise attention weights by matrix multiplications and element-wise operations, and then applies them to the feature maps. Additionally, we replace the SPPF module with a context augmentation module to improve feature map resolution and quality. To minimize information loss, we utilize Carafe upsampling instead of the conventional upsampling operations. Furthermore, we use a loss function that combines IoU, binary cross-entropy, and focal loss to improve bounding box regression and object classification. We use stochastic gradient descent (SGD) with momentum and weight decay to update the parameters of our model. Through rigorous experimental validation, our S-YOLO model demonstrates outstanding accuracy and efficiency in weld defect detection. It effectively tackles the challenges of small target detection, target occlusion, and target overlap. Notably, the proposed model achieves an impressive 8.9% improvement in mean Average Precision (mAP) compared to the native model.

More from our Archive

• DOI: 10.3390/axioms12070659 2023

  Study of Thermoelectric Responses of a Conductive Semi-Solid Surface to Variable Thermal Shock in the Context of the Moore–Gibson–Thompson Thermoelasticity

  Sami F. Megahid, Ahmed E. Abouelregal, Sameh S. Askar, Marin Marin

• DOI: 10.3390/axioms12070697 2023

  A Novel Weld-Seam Defect Detection Algorithm Based on the S-YOLO Model

  Yi Zhang, Qingjian Ni

• DOI: 10.3390/axioms12070675 2023

  A Unified Approach for Extremal General Exponential Multiplicative Zagreb Indices

  Rashad Ismail, Muhammad Azeem, Yilun Shang, Muhammad Imran, Ali Ahmad

• DOI: 10.3390/axioms12070658 2023

  New Conditions for Testing the Asymptotic Behavior of Solutions of Odd-Order Neutral Differential Equations with Multiple Delays

  Fahd Masood, Osama Moaaz, Sameh S. Askar, Ahmad Alshamrani

• DOI: 10.3390/axioms12070691 2023

  Some New Bullen-Type Inequalities Obtained via Fractional Integral Operators

  Asfand Fahad, Saad Ihsaan Butt, Bahtiyar Bayraktar, Mehran Anwar, Yuanheng Wang

• DOI: 10.3390/axioms12070694 2023

  Exact and Approximate Solutions for Linear and Nonlinear Partial Differential Equations via Laplace Residual Power Series Method

  Haneen Khresat, Ahmad El-Ajou, Shrideh Al-Omari, Sharifah E. Alhazmi, Moa’ath N. Oqielat

• DOI: 10.3390/axioms12070680 2023

  On Estimation of Reliability Functions for the Extended Rayleigh Distribution under Progressive First-Failure Censoring Model

  Mahmoud Hamed Abu-Moussa, Najwan Alsadat, Ali Sharawy

• DOI: 10.3390/axioms12070667 2023

  Infinitesimal Affine Transformations and Mutual Curvatures on Statistical Manifolds and Their Tangent Bundles

  Esmaeil Peyghan, Davood Seifipour, Ion Mihai

• DOI: 10.3390/axioms12070665 2023

  Determinantal Expressions, Identities, Concavity, Maclaurin Power Series Expansions for van der Pol Numbers, Bernoulli Numbers, and Cotangent

  Zhen-Ying Sun, Bai-Ni Guo, Feng Qi

• DOI: 10.1142/s0218348x23500597 2023

  LOCAL GEOMETRY OF SELF-SIMILAR SETS: TYPICAL BALLS, TANGENT MEASURES AND ASYMPTOTIC SPECTRA

  MANUEL MORÁN, MARTA LLORENTE, MARíA EUGENIA MERA