Major League Hacking's Local Hack Day @ UTSA
A 12-hour hackaton across several countries and schools. More info on Local hack day's website
Pet Finder: a web app for pet face detection.
Surveillance cameras are now able to use AI software to identify individuals on public places. The motivation behind this project is to use this technology to find missing pets.
The Google Cloud vision API was used to quickly provide a description of the contents in an image. No need to train my own neural network!!! 😅 On the back-end, NodeJS was used to create the simple web app. Bootstrap was also used for some quick formatting.
An image of a pet is sent through a request to the API, and a response is received containing a description (e.g. "dog" or "cat") and a score value from 0 to 1, which indicates the level of confidence (1 = 100% confident) for each description.
The app is running with 3 sample images. It sends the request to the vision API and displays the descriptions and confidence levels as a list for each image.
- Need to add an HTML form or other means to allow the user to upload their own images.
- In order to recognize a particular pet, a database of pets would be necessary. I.e. the problem is not just recognizing an animal in a picture, but recognizing a specific animal/pet.
- Would be nice to also make it a mobile app.