Published Paper Details:

Defect Detection in modern material assembling is a fundamental requirement for which productive practical arrangements must be sought. In ongoing years, a few distinct techniques have been created to recognize these deformities, some of which have been fruitful to a more prominent or lesser degree. Traditionally, the acknowledgment of texture has a lot of difficulties because of its manual visual examination. Additionally, the methodologies dependent on early AI calculations legitimately rely upon high-quality highlights, which are tedious and blunder inclined cycles. Hence, an automated system is needed for the classification of fabric to improve productivity. In this paper, we propose a solution in which a deep transfer learning model will be trained on a fabric dataset. The training process will include various data augmentation techniques like rescaling, zooming, horizontal flipping, etc. The model extracts important features from a given image and classifies automatically in an end-to-end fashion. We assessed the consequences of our model utilizing assessment measurements, for example, exactness, adjusted precision,AUC and Recall.