The presentation focuses on the use of computer vision for quality control in manufacturing, particularly in the textile and clothing industries. It begins by highlighting the limitations of traditional manual inspection, such as human fatigue, subjectivity, and inefficiency, and motivates the need for automated defect detection systems. Computer vision is a robust and flexible solution capable of analyzing material properties like texture, shape, and color. The presentation then details an algorithmic approach for defect detection using a bank of Gabor filters. The method involves dividing images into regions, applying multiple filters, and selecting the most discriminative one using a cost function based on output variation. This enables segmentation of defects such as stains or yarn issues. However, the limitations of manual filter selection lead to the introduction of deep learning approaches, particularly Convolutional Neural Networks (CNNs), which automatically learn optimal features from data. CNNs provide advantages like hierarchical feature extraction, spatial invariance, and end-to-end learning, significantly improving detection accuracy and reducing reliance on expert tuning. Finally, the presentation showcases practical implementations, including a defect detection system for airbag fabric manufacturing and sewing quality inspection. It describes the evolution from single-camera setups to multi-camera systems with dedicated analytics servers, achieving measurable improvements in defect reduction and maintenance efficiency. Additional applications include automated sewing defect detection, as well as chip geometry measurement in machining processes using high-speed imaging and advanced image processing pipelines. The conclusion emphasizes that computer vision enables faster, more reliable, and scalable quality assurance across production stages, supporting better control, reduced costs, and improved product reliability.