Breeder App

[Jewels]

I found an app that will identify plants. I plugged in Costa Dave's photo and it id'd correctly. I was curious if software could be trained to help identify pairing
results by leaf shape.
Like if you were to take 20 photos of a mother fan leaf, and 20 photos of a father fanl ea ,, could be app be trained to identify features in the offspring ?

 

[Frimpong]

I found an app that will identify plants. I plugged in Costa Dave's photo and it id'd correctly. I was curious if software could be trained to help identify pairing
results by leaf shape.
Like if you were to take 20 photos of a mother fan leaf, and 20 photos of a father fanl ea ,, could be app be trained to identify features in the offspring ? View attachment 31842

Very cool !

 

[Sammy Snakeskins]

I like it. There's an app I used for my veras(not aloe veras but I forget the name). You submit your pics to the app support botanists and they identify the plant for you. First two identifications are free I think then you pay a $1 for any other plant you want identified.

 

[Dino Party]

Cool concept but with everybody and their grandma and even me chucking today, I cant imagine that it would be very reliable for cannabis, other than the basic thin leaf/broad leaf identifications.

 

[Gentlemancorpse]

Cool concept but with everybody and their grandma and even me chucking today, I cant imagine that it would be very reliable for cannabis, other than the basic thin leaf/broad leaf identifications.

That's kind of what I was thinking... probably not gonna ID Joe schmoe chuckers Animal Cookies x Purple Elephant x Thors Hammer shenanigans... still a really cool idea though!

 

[Jewels]

First two identifications are free I think then you pay a $1 for any other plant you want identified

This app is automated and free.

.. probably not gonna ID Joe schmoe chuckers Animal

This current app could not. But what about learning software ?
There are dudes on here who can spot a thrice worked mother. Seeing Moms features , even a generation or two later.

Facial recognition software can beat a beard. I don't think this idea is a huge stretch.

I just want a "session trainable" identifier.
For example.
I submit 20 photos of a lion.
20 photos of a tiger.
Then I give the AI a picture of a crossbreed.
I want the app to reply:
" Liger - ( Lion x Tiger)
This sample is 65% Lion "

 

[spyralout]

kinda like google image search for plants. this would be really cool. it could also help w spotting fakes.

 

[HydroRed]

kinda like google image search for plants. this would be really cool. it could also help w spotting fakes.

or spotting lizards in your canopy

 

[Gentlemancorpse]

This app is automated and free.

This current app could not. But what about learning software ?
There are dudes on here who can spot a thrice worked mother. Seeing Moms features , even a generation or two later.

Facial recognition software can beat a beard. I don't think this idea is a huge stretch.

I just want a "session trainable" identifier.
For example.
I submit 20 photos of a lion.
20 photos of a tiger.
Then I give the AI a picture of a crossbreed.
I want the app to reply:
" Liger - ( Lion x Tiger)
This sample is 65% Lion "

It's certainly possible someday but it would take quite the advanced AI... the Liger example is far simpler than it would be with cannabis.... it's two different distinct species that haven't been selectively bred, with distinct phenotypical differences... with cannabis it's 3 species that have been selectively bred for thousands of years (and honestly, it's really one species as far as I'm concerned being that the actual definition of a species is a group of organisms that can breed and produce viable offspring... obviously we can cross a sativa and an Indica then cross those plants offspring and still make seeds, whereas all ligers are sterile)... so the differences between strains are far more nuanced... it's theoretically possible to have two strains with nearly identical physical phenotypes and different terps... which obviously can't be distinguished from a picture...experienced growers get around this by using deductive reasoning... if you know the grower, or breeder or even just current trends in growing you can use visual cues and that other knowledge to make what's essentially an educated guess. Obviously you can teach an AI that stuff too but it becomes a staggering amount of data

Now if someone invents smell-a-vision all bets are off hahaha, and obviously, I'm not an AI expert, I'm just spittballing here

 

[jaguarlax]

Cool concept but with everybody and their grandma and even me chucking today, I cant imagine that it would be very reliable for cannabis, other than the basic thin leaf/broad leaf identifications.

We could make it reliable for pH stuff perhaps with enough leg work...