Photorealistic Image Filters

Cyanapse's AI-powered image filters enable state-of-the-art image transformations while preserving fine-grained image features

Photorealistic Enhancements with AI

Automatically enhance your photos and give them the look you want with AI-powered filters

Photorealistic Image Filters

Cyanapse's Photorealistic Image Filters are based on a cutting-edge method that leverages recent advances in AI to transform and give new looks to images in just a couple of seconds. Our implementation allows to automatically edit photos with various photorealistic transformations, including changing the night into day or day into night, adding golden hour photography effects to increase warmth, enhancing the lighting to reflect natural light conditions, and much more.

Day

Before
Bodegas - Before

 

 
Double Arrow

Night

After
Bodegas - After

Cyanapse's AI-enhanced filters enable photorealistic editing. The filters are based on neural networks that have learned how image textures look like under different lighting conditions, and can be applied on any image to imitate different lighting effects on all kind of textures. Our technology also enables the use of bespoke photo styles that can apply a brand's visual signature on any photo.

Natural light effects are based on an AI technique that is similar to artistic style transfer; a method that is used for complex transformations such as applying an artist’s brush style to given images, or to generate images from doodles. It is based on a deep neural network that “learns” the textures of objects in the images that it had seen, which it then applies to any given image based on the features that it has recognised in the image to be transformed. Using this method, AI ‘imagines’ how the image would look like under different conditions, an thus imitates natural light effects.
Unlike artistic style transfer, photorealistic translations from one natural image domain to another requires strict reliability to the image content, which is a challenging task as the features that can be encountered in casual photography can be various. In order to develop photorealistic effects for automated image transformations, we use Generative Adversarial Networks (GANs) with a number of improvements to enforce picture quality. Our methods include constraints that ensure photorealism and conservation of the image content.

AI for Image Editing

Photorealistic image filters will take your photo editing skills to the next level

Image Editing Workshop

Whether you are a photographer, a blogger or a content developer, automating your image editing tasks will save you time.

Our enhancement techniques will enable you to visualise your images with different styles at a glance, and help you achieve professional-level photo editing in a blink.

Different Styles for Your Photos

Touch the images below to view image enhancements in high resolution

Enhancement

Get the most out of your photos with our dramatic effects.
Before
West Pier - Before
After
West Pier- After

Day ↔ Night

Turn day images into night images (and vice versa).
Before
Istanbul Bosphorus - Before
After
Istanbul Bosphorus - After

Brightening

Brighten up your pictures as if they were taken under plain daylight.
Before
Brighton Pier - Before
After
Brighton Pier - After

Golden Hour

Add more warmth by imitating golden hour photography.
Before
Clouds - Before
After
Clouds - After

Try it Yourself

It works with all kind of photos, check it out below with these example images and pre-generated styles

Image: Albion

Image: Brighton Pier

Image: Downs

Image: Istanbul

Looking for NVIDIA GTC EU 2018 (Munich) high-resolution images? Click here.

Learn More & Early Access

Get in touch to learn more about our photorealistic image filters
and get a free beta access to Celeste Photo App and Cylight Vision API

Your personal information is safe with us and we respect your choices. We will not share your details with third parties unless we have your clear consent to do so, and you can delete your data from our systems at any time using the 'Manage Subscription' button at the bottom of every page. Please check our GDPR-compliant Privacy Policy for more details.