BLUEDOT: AI-Based video super resolution architecture and performance improvement using Catapult HLS

In recent years, high-resolution image generation technologies—particularly Super Resolution (SR)—have attracted significant attention. SR is a technique that converts low-resolution images into high-resolution ones and has been applied across various domains. With the introduction of deep learning techniques to traditional interpolation-based methods, SR performance has improved dramatically.

However, the performance of deep learning-based SR has proven insufficient for video processing, and adapting such techniques to meet customer-specific requirements has become a time-consuming and costly endeavor. In this context, the adoption of Catapult High Level Synthesis (HLS) has provided a breakthrough solution.