Neural Networks in Russia to Help Find Lost Pets by Whiskers, Paws and Tails
On Sakhalin, specialists are training neural networks to analyze images for monitoring urban infrastructure
Specialists at the Sakhalin Artificial Intelligence Center are training algorithms to analyze photographs and assist municipal services. Women programmers are also involved in developing the technologies.
Neural Networks Serving the City
At the Sakhalin Artificial Intelligence Center, computer vision algorithms analyze images to assess the condition of urban infrastructure.
The neural network can now determine whether snow has been cleared from building roofs, detect ice accumulation or identify blockages in spawning rivers. The algorithms are also used for environmental monitoring. For example, the system can track the spread of dangerous plants.
Searching for Lost Animals by Photo
One of the center’s developers, Polina Vitkovskaya, is working on a Telegram bot designed to help locate lost pets. Users can upload a photo of a missing animal, and the service compares it with an existing image database.
The algorithm identifies distinctive features of the animal and searches for similar photos online. Pet owners then receive clues about where announcements of found cats or dogs with similar appearances have been posted.
Hogweed vs. Algorithm
Another project is led by Lyubov Shumilova, a specialist at the center. She trained a neural network to recognize the invasive plant borshchevik (hogweed) in photographs.
Using image analysis, the algorithm generates a map showing the spread of the hazardous weed. The data can then be used to plan eradication efforts.
Women Developing AI
Today women occupy about one-third of positions at the Sakhalin Artificial Intelligence Center. They develop algorithms, process data and build new digital services.
