A COMBINED APPROACH TO IMPROVING IMAGE QUALITY USING BICUBIC INTERPOLATION AND A L ANCZOS FILTER

Abstract

In today's digital world, image processing plays a key role in numerous industries. From medical imaging, which requires high definition for accurate diagnoses, to satellite reconnaissance, which relies on detailed images to monitor terrestrial and atmospheric phenomena. This also applies to multimedia products, where image quality directly affects the user's perception of the content. One of the main challenges in this area is to increase the size of images while maintaining their quality. Traditional methods, such as bicubic interpolation and the Lanczos method, have long been used in this area because of their simplicity and efficiency. While they are effective for moderate image enlargement, when enlarged significantly, they often result in quality loss, such as blurring and poor detail preservation. This is especially critical in situations where high detail is required, such as medical images for accurate diagnosis or satellite images for precise monitoring of changes on Earth. Therefore, there is a need for more advanced methods of image enlargement that can preserve high quality details even at high magnification. In response to these challenges, modern technologies, such as deep learning and neural networks, offer new approaches to image processing. These methods can significantly improve the quality of images, in particular by enhancing the reproduction of details and reducing blurring. However, they require significant computational resources and large datasets for effective training, which can be a limiting factor for their application in practical scenarios. In addition, the complexity of setting up and fine-tuning neural networks can also be a barrier to the widespread use of these methods. Given these limitations, a combined approach that includes elements of classical techniques seems to be the best solution. This method can combine the simplicity and reliability of classical methods with high quality and detail. This allows achieving high image quality with moderate computational costs. This article analyzes and compares classical methods in detail with the combined approach, using various quality metrics such as PSNR to objectively evaluate the results. In addition to image quality, processing time and resource efficiency are also considered, which are important aspects in choosing an image processing method. A combined approach that integrates classical and state-of-the-art methods can be an effective tool in modern image processing applications, providing high quality and efficiency.