ColouriseSG: the free application with which you can convert your photos in black and white into colors

A group of application developers presented an online tool that can instantly bring black and white photographs to life by adding color. by artificial intelligence.

The colorization of old images is a process that usually requires a lot of time. and you need specialized training and expensive software.

L & # 39; tool ColouriseSG, You can do it for free from a single digital image and work with iconic historical photographs and old family portraits.

The application is formed in a subsequent catalog of old images and uses machine learning to guess what the color image would look like.

"The purpose of colorization is to generate an image with plausible colors that can coexist. In no way guarantees that the colored image is an accurate representation of the instant photo in time", said the developers of the tool.

The application is also more able to color images of human subjects with natural landscapes and can fight with more complex images. The free use tool has been developed to offer Singaporeans a precise way to edit their monochrome images.

Existing software, such as Algorithmia, is trained with the aid of 1,300,000 images from ImageNet., a database of photographs developed in the United States by researchers at Stanford University and Princeton University.

ColouriseSG's idea was to provide a sufficiently large dataset to be relevant to Singapore residents.. For its part, technology company NVIDIA has launched several different image editing tools that work with artificial intelligence.

Its latest version is a software that uses in-depth learning to turn even the most difficult images into works of art.. The new program, called GauGAN, uses a tool called Generative Confrontation Networks (GAN) to interpret simple lines and turn them into hyperrealistic images.

Its application could help professionals in a wide range of disciplines such as architecture and urban planning make images and visualizations faster and more accurate, according to the company.

"It's much easier to create ideas with simple sketches, and this technology can turn sketches into very realistic imagessaid Bryan Catanzaro, vice president of learning research at NVIDIA.

What separates GauGAN from other rendering software, besides the simplicity of its interface and its application, lies in the development of this tool.

According to NVIDIA, GauGAN has been trained to emulate different types of landscapes with the help of an in depth learning software allowing to study more than one million different images..