Toeshoeman last edited by
I use my MM2 in my bedroom and I use it every morning. It would be beneficial for me to look at the mirror and have an outfit idea up for me.
- Summary: In between the hours of 5:30am to 7:30am, display the outfit idea somewhere on the screen. It can then be cleared using a hand gesture, or it can remain on the mirror.
- Setup and Logic: The user will have to take a picture of all of their articles of clothing and label them something basic (i.e. Red Long Sleeve Shirt.jpg, Blue Jeans.png). The module will read the forecast for the day and select based upon the high temperature and precipitation. The user will set their temperature threshold (40F to 55F = long sleeve or 75F to 95F = shorts). The module will select from different buckets: shorts, pants, long sleeve, short sleeve, sweater, jacket, collared shirt, button down, etc. There will be basic logic to assist with matching: if red shirt, then no red pants.
It is a rough idea, but I would love to see it come to fruition. I am more than happy to work with someone to code this. I do not have experience coding modules but I do have experience with Fortran (stone age stuff).
plumcraft last edited by
I had the same idea
I think issues are :
- How to store these data ? Into a database ?
- Take pictures of all your clothes (which take a long time)
Deferring on the picture taking as that’s just a painstakingly long process depending on how many articles of clothing you want to store. And how, whether you’re going to wear each one, take a picture, and then clean it up, or if you’re just hanging them on some background. How are you going to do ties? Again, wearing each and cleaning up the image to only show the tie? Yeah, deferring.
The better practice is to store (TEXT) data like that in database tables, and create indexes that can then be quickly searched. For example, have a table with only the shirts, listing color, style, and whatever other attributes you deem necessary. Do the same for pants. Have both indexed. Then you need to create some logic that will determine what would match based on your preferences, so you don’t end up wearing colors that clash, for example (then again, maybe you are that stylist who does exactly that.)
It’s not impossible, it just takes a lot of work. The pictures part, as well as the logic part. It’s really easy to query the database for a random shirt and pants. It’s harder to query for something that makes sense.
Toeshoeman last edited by
@plumcraft @KirAsh4 I agree, taking pictures would be painstaking. That is a good idea to do text data – like a description to that shirt, and then matching that with pants. I’m also talking super basic here: shoes, pants, shirt, casual jacket. Do you think you could help me with a skeleton on that?
martin last edited by
I wanted to let you know that I’ve built a service that does everything you require… except for the actual smart mirror bit…
I’ve just posted in the General Discussion part of the forum so there’s more info there but in short, the service is called Able Style and it’s a virtual personal stylist service. It’s already availble as a voice service on Alexa and Google Assistant so it may already work with the smart mirror using those services as a conduit.
Would love to know what you think.