PhotoGuru switched to the "dark side" and "wiser"

    I continued experimenting with a prototype of PhotoGuru to see if recognizing the "style" of a photo can help the user in some way.

    In the new version, the interface is updated and the program gives only one answer. Most often, this is the name of the photographer and the percentage of “similarity” of a user's photo to the style of one of six photographers.

    If the percentage is lower than 33%, the system cannot establish the author and concludes that in front of her is a new “guru” and instead of the last name shows the inscription “PHOTO GURU”.

    if (maxPercent <= 100/numPhotographers*2) {
            message = " ★ "
            photoGuru = "Photo Guru"

    I added my photos and portraits of Zhenya Otetskoy to the system and the model began to make mistakes in 31% of cases.
    Having edited the samples, I realized that I was starting to adjust the answers to the result. Need another way. It turned out that the system can be taught longer. I increased the number of training cycles to 20 and now the system is almost not mistaken when it analyzes photos from a test sample.

    In general, the model correctly classifies portraits unknown to her, but the system considered one of my old photographs to be a photograph of Zhenya Otetskoy and I'm not quite sure that I myself was shooting a self-portrait.

    But most of all I am impressed that by training only on portraits, the system was able to confirm with 99% accuracy that the landscapes with the “trains” were shot by Eugene Otetskiy.

    I tried to find a photograph that would not be similar to any of the authors known to the system. I could not get.

    The system considered that the photos from Tim Walker are similar to the photos of Zhenya Otetskoy, Mike Brodie shoots like Egor Voynov, the abstraction was made by Maxim Shumilin.

    Most likely, the model analyzes photographs at the composition level and therefore produces such a result. She cannot appreciate aesthetics and symbols, and she recognizes the white square as a photograph of one of the authors.

    Playing with PhotoGuru can be useful in self-learning photography. To do this, you need to increase the number of photographers known to the system, think about protection from cheaters and add “rewards”. For example, the “Videnin 99%” badge and the “big medal” to the user who collected the “99%” badges from all photographers.

    And if someone gets the status of “PHOTO GURU” - a diploma for uniqueness.

    Also popular now: