I take a lot of photos with my phone. About every month I back them up to my computer and sort them into folders. Even though the photos most contain my pets, I have other photos like of my motorcycle, receipts, funny Reddit posts and funny conversations. Here's an example folder layout:
- Pictures
- Jane
- Ginger
- Stanley
- Huginn
- Cat Combo
- LG G6
- Conversation
I can use machine learning to sort through these photos and organize them a heck of a lot quicker. Right now, the machine learning is using AWS' database. This means that it may not label my photos correctly. For example, one photo of the dog on a floral bed labeled the photo as "floral" rather than "dog." I'd like to implement a cross-check with Google Vision.
From this small weekend project, I learned more about Python, functions, argument parameters, for loops, closing files properly, and closing GitHub issues.
Eventually, I will figure out Tensorflow and manually train a model to sort for me. But this is a great first step for me.
Check out the GitHub repo here: