About Project

This project is dedicated to the stray dogs of Batumi, Georgia.

The goal of the project is to demonstrate the capabilities of neural networks in addressing the issue of tracking stray dogs through image recognition. Here, photos of dogs taken by volunteers on the streets of Batumi using the Telegram bot @batumidogs_bot are collected.

When a photo is sent to the Telegram bot, it is processed by a neural network. The program recognizes the dog's face in the image, then compares it with other previously added photos of dogs in the database. If a similar dog is found in the database, the photo is simply added to that dog's record. If the neural network does not find a similar dog, a new entry is created in the database.

You can add your own photos of dogs in Batumi. To do this, find the Telegram bot @batumidogs_bot. Send photos to the bot where the dog's face is clearly visible in a frontal view, not in profile. The search for dogs is based specifically on frontal shots of their faces, and both eyes should be visible. The bot will either return an image of a dog found in the database with the phrase "The dog is found," or it will respond with "The dog is added." The photos will appear on this website.

The neural network is still in the learning and testing phase, so errors may occur.