You can find artificial intelligence that generates fresh photographs, but what if you want to repair an old family photo? You may have a no-cost alternative.
Free AI Tools
Louis Bouchard and PetaPixel have brought attention to GFP-GAN (Generative Facial Prior-Generative Adversarial Network), a free tool recently developed by Tencent researchers that can recover damaged and low-resolution photographs. The system combines information from two AI models to fill in a photo’s missing details with realistic detail in a matter of seconds while retaining high precision and quality.
In order to restore images, conventional methods fine-tune an existing AI model by comparing artificial and real images. According to scientists, this frequently produces subpar outcomes. The novel method employs a pre-trained version of an existing model (NVIDIA’s StyleGAN-2) to inform the team’s own model at numerous phases of the picture-generating process. The technique seeks to preserve the “identity” of people in photographs, with an emphasis on facial characteristics such as eyes and mouths.
This endeavour is still constrained by the limitations of contemporary AI. While remarkably accurate, it makes educated estimates about what content is missing. The researchers cautioned that you may observe a “slight alteration of identification” and a poorer resolution than desired. Folks, do not rely on this to print a large photograph of your grandparents. Nonetheless, the work presented here is intriguing; it suggests a future in which it will be simple to recover photos that would otherwise be lost to the ravages of time.