Is there a way to modify the label.txt file?

[mking1011]

Is there a way to modify the label.txt file?

I just changed the label.txt using editplus.

Is there a problem with my method?

[Burgomehl]

Of course you can change a txt file like this, but what do you think will be the result of it?

[mking1011]

@Burgomehl The results were not good. Can you tell me another good way?

[Burgomehl]

Just changing the label.txt file will not show you some new detected objects. You need to train the model you are using.

Or what do you expect it to do?

[mking1011]

@Burgomehl Ah! I made the question strange. I have a newly trained model. I would like to have the appropriate label.txt

[Burgomehl]

(From the Readme) To use your own model:

Make sure your model trained with TensorFlow 1.4 if you use 0.3 version of the Unity plugin that I linked above. You can also try 0.4 version that uses TensorFlow 1.7.1. Change extension of your model from .pb to .bytes. Put your model and labels in Resources. Set Model file and Labels file to your model and labels in main camera object of the scene you chose. If neccesary, change classifyImageSize, IMAGE_MEAN, IMAGE_STD, INPUT_NAME and OUTPUT_NAME to suit your model.

Why was the Result not good?

[mking1011]

@Burgomehl I trained two classes. But only one class is well aware

[Burgomehl]

I think it's all about four fist and palm object detection? (https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10/issues/148) How did you train? How many images? What does your config file look like?

[mking1011]

@Burgomehl Here are the files I trained (EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10#148)

I trained with 16000 fist photos and 16000 palm images

[Burgomehl]

You found a solution? @mking1011

[mking1011]

@Burgomehl I increased the number of pictures with fists and palm. That solved the problem.