The PhD focuses on developing a modelling framework for classification and defect recognition. Data made available consists of ultrasonic and photographic images from different applications. The goal is to develop a data-driven modelling framework for constructing models capable of providing linguistic explanation for their classification results.
Linguistic explanation can assist a non-destructive testing (NDT) expert in understanding a classification result further than results with model-based explanation. A range of machine learning techniques will be used that can provide interpretability and linguistic output.