Image selection technology

ReSnap helps you pick the best photos and filter out the similar ones. We create the first 80% of your photo book and the last 20% you easily personalise yourself.

The images are selected by analyzing all the images step by step. For instance, we filter out photos by looking at several quality aspects (like blurriness). Our system also automatically looks at the amount of faces and the micro expressions of the faces. Of course we filter out similar photos by looking at the pixels. After this part our deep learning network will look at correlations the pictures in your uploaded photo set. In this way we analyse what kind of picture you would like to have in your photo book. So basically, we try to predict the selection of images you would make. This algorithm is continuously updated and improved, and will be getting better and better in the future. So keep supporting us!

ReSnap Image selection technology

  • Our algorithms enable fine grained classification of objects and quality assessment of images on all images, which are valuable for you as a user. We’ve developed high quality and super scalable neural networks to understand your photos.
  • We’ve developed high quality and super scalable neural networks to enable the understanding of the emotional value of your photos, and similarity search based on the combination of semantic and visual properties.
  • After the analytics part, which is done in under 100ms per photo (much faster than human annotators), our system ranks your photo set on emotional value by taking into account more than a million different parameters.
  • Our auto layout algorithms place the selected photos perfectly into a photo album. It always calculates the most efficient way to fill up your photo book chronologically.

You probably will recognize the following situation. You want to take a picture of an old church on a square. The first one you take is a bit overexposed, so you change the focus. It doesn’t work, so you change the angle, but now you miss a part of the square. You move a bit, but now the picture is blurred and you take another one. Finally, after ten attempts, you took that good photo. You don’t remove the bad ones, because this takes too much time and you want to enjoy your view. Imagine that this happens 50 times during a holiday, you can easily have 450 of pictures that are similar or have a bad quality. Our neural network will understand your problem and it will filter out the photos you don’t want to keep. Why still spending time organizing your pictures if our technology can do that for you?