TMIC: App Inventor Extension for the Deployment of Image Classification Models Exported from Teachable Machine

is it only available the image classification? how about the pose object?

Hi, the extension only runs image classification. We are currently not planning an extension for pose detection.

We are trying to create a classifier for different types of fish. But it is always resulting in the same classifier result, no matter which picture we show it. If trying on Teachable Machine directly, it seems to show different results, but when in AppInventor, it is always giving just one result.

Hi,

I tried using the example, but i ran into this problem with the runtime error. what should i do?

If you show result in a Label, we can see what is in it.

thx. That worked for me

Hi. I've used the TM extension with the associated web viewer and it generally works, but I would implement a different way:

  1. I have a button to call ClassifyVideoData and this procedure is fine, but I would like to freeze the image in the web viewer when I press the button to take a photo, instead of having a continuous video in the web viewer
  2. I would also like to select a photo from the smartphone gallery using the Imageselector button and the ClassifyImageData block, having previously enabled the extension's InputMode in "Image". But it doesn't work!! It seems like it keeps decoding what it sees on the camera!
    How can fix this?
    Tx a lot!

hello can you help me, i have this error


If specific key lookups in the result dictionary return "not found", the next step in diagnosis would be to list the keys of the returned dictionary and see how they differ from your expectations.

so how can i correct this error

set result1. text to result.
and remove other two from the got classification event.

Replace "not found" by 0 to be able to round correctly in case the key is not found

Taifun

thank you can you give an idea if i want to change take picture with camera i want to upload picture from files on phone

can i attach teachable machine with picture uploaded from file

hello i need help about teachable machine image classifier i want to use picture uploaded from file not from camera thank for advance

Hello, this version of the extension does not provide support for the upload of images, only for images taken with the camera

Hi,
I have used this extension in the past and it worked wonderfuly. I was wondering if you could consider to include a toggle camera feature as is used for example in the LookExtension.

Kind regards