Artificial Intelligence (AI) has had a massive impact across various sectors, especially in the image editing/ generating segment, as tech companies have developed various advanced AI-based tools to enhance and generate images. And now, a recent entrant is an AI-based image-enhancing tool that promises to generate more accurate facial features of individuals in portrait images than previous AI-based imaging tools.
Dubbed as GFP-GAN, which is short for Generative Face Priority based on the Generative Adversarial Network architecture, has been developed by researchers at the Tencent ARC Lab in China. In a recent official paper, the researchers detailed their latest AI-based image-enhancement tool that uses the GAN architecture to restore faces of individuals in old, blurry, and damaged images, and generate an upscaled version of it.
The difference between GFP-GAN and other AI-based imaging tools is that it prioritizes the restoration of the facial details of individuals while reducing unwanted artifacts. The team of researchers also created a demo of the tool to showcase its feature.
“While previous methods struggle to restore faithful facial details or retain face identity, our proposed GFP-GAN achieves a good balance of realness and fidelity with much fewer artifacts. In addition, the powerful generative facial prior allows us to perform restoration and color enhancement jointly,” the researchers wrote in the paper.
As per our testing, the GFP-GAN tool works like a charm for blurry and old images, providing you with a high-resolution image, however, generates some creepy-looking images.
Furthermore, the app completely crashed, displaying a “runtime error”, after we used it on an image with multiple faces. Although the CPU memory limit did not exceed, it showed the same repeatedly.
Plus, critics suggest that AI-based image-enhancing tools like GFP-GAN can pose a privacy threat for many as these can be used to monitor populations and improve CCTV images of people without their permission. Nonetheless, it is not as bad as the Clearview AI system, I guess.
What do you think about GFP-GAN? Do you think these tools can benefit society in the future? Let us know your thoughts in the comments below.