The idea of winter-to-summer image translation appeared to be quite interesting and challenging. Numerous difficulties can arise during this process in terms of dataset aggregation and preparation; long iterations, the cost of experiments; the complexity of generative models, the configuration of many knobs.

However, our R&D engineers have learnt how to make it with just one click. After the comprehensive research of current possibilities of Machine Learning generators and what quality they can provide, our specialists prepared a solid dataset, and adopted the CGAN (Conditional Generative Adversarial Networks) to satisfy our requirements.

Although there are a lot of research works and testing of image-to-image translation, not a lot of specialists are ready to share their promising results. Our experts are happy to show their demo to you. Less talk, more actions: try out the algorithm and share your feedback with us!

Features

  • Sharp & realistic image synthesis.

    Output images look realistic and seem to be almost the same as the ground truth images.

  • Possible implementation in landscape design and related spheres.

    It gives an opportunity to have a clear vision of the same image in different seasons. The implementation of the same technologies can also enable the various color changes of images.

  • High algorithm recognition speed.

    The algorithm can process frames from your camera in real time.

  • Works on low resolution images.

    The algorithm is able to translate winter image into summer image even on low resolution images.

How it works?

how it works mobile image

Try it yourself

Choose one of the sample images and get the result.

Result

Input Image
sample
Output Image
sample
Ground Truth
sample

Contact Us

To find out more about Abto Software expertise, request a quote or

get a demo of your custom solution.