Nowadays the use of radiography to inspect weld defects is of great importance in different industries. Given the various causes for image quality reduction in radiography systems, the use of image processing to enhance the contrast of radiographs is crucial. Artificial Intelligence (AI), as one of the most advanced technologies of the modern era, plays a significant role in the image processing, where machine learning and deep learning algorithms are employed to analyze and interpret visual data. In this research, Facebook AI Similarity Search (FAISS) has been used to improve the contrast of weld radiographs. FAISS is a powerful and optimized library for similarity search in large datasets, developed by Facebook. The results of processing the radiographs show that the contrast has increased in various regions, particularly in the weld root and defect areas, where gas porosity and lack of fusion are most prevalent, showing a significant improvement. These results have been evaluated by radiography experts, who confirm that, in addition to improving the contrast of radiographs in different regions, the defect detection can be carried out efficiently. In addition, this method is fast and does not require complex manual adjustments. One of the key advantages of this method is the use of a pre-trained network, which saves time and costs associated with training new models. This is particularly important in large industries such as oil and gas, where time and accuracy of detection are critical. Given the positive results of this research, it is expected that the use of AI and libraries like FAISS will become a standard tool in the processing and analysis of radiography images in the future, bringing about a fundamental transformation in the quality and speed of defect detection, and it can help to identify the weld defects and discontinuities by radiography and welding specialists.