Aselsan has successfully demonstrated the real-time resolution upscaling prototype capability of a device developed to generate high-resolution images from low-resolution sensors.
As part of the prototype capability demonstration, images captured from a camera connected to the graphics processing board were processed in real-time by a deep learning-based resolution enhancement algorithm, increasing the resolution fourfold. The underlying algorithm is capable of increasing the resolution of both TV and thermal images by two, four, and eight times. In this context, pixelation in the images was eliminated, the outlines of objects in the image were made sharper, and illegible text was made legible.
Another point worth highlighting is that despite the low power consumption of the graphics processing card (15W), the resolution could be increased in real-time and with virtually no lag.
