Bengaluru Girl Develops AI Tool To Recognise & Track Stray Dogs For Vaccination

Stray dogs are certainly a problem in India, as even animal lovers can attest to.

Thanks to a combination of careless breeding, abandonment, and a lack of sterilization programmes, there are more of them in most cities than can be effectively handled.

But you can't just get rid of them, that would be terrible. If you want to keep them healthy though, and ensure they're not a harm to the human population nearby, you need to ensure they're vaccinated. So an ingenious student from Bengaluru decided she would adopt artificial intelligence to ensure that kind of initiative can be efficiently carried out.

Aparna Ajit Gupte, a grade 12 student of National Public School in Indiranagar, says she was drawn to the problem thanks to all the news articles she read about human deaths caused by rabid stray dogs. She found that the vaccination attempts by civic authorities could be made much more efficient if they had some sort of recording system. So she figured she'd just build one.

Gupte first set about obtaining pre-trained neural networks capable of identifying dogs in images extracted from video recordings. She then used a face recognition algorithm and modified it to instead detect unique markings and features of stray dogs. Then, all she had to do was classify them.

With her model, Gupte was able to obtain a 92 percent accuracy when it came to individually identifying the different dogs in a locality. To use it, all you have to do is keep updating the data you're working with by adding images of the dogs vaccinated. The next time, it'll be much easier to pick out the ones not tackled yet.

In fact, Gupte even suggests this system could be automated, by tracking the dogs using footage from street CCTV cameras. Her plan in the future is to track the dog population in an area as it ages by continuously updating the dataset with new images.