Hi @Yves Guena ,
Thank you for using Microsoft Q&A.
Form Recognizer does not have an inbuilt function to convert bounding boxes from a labels.json file to pixel values. However, you can use the formula pixel_value = bounding_box_value * image_size
to calculate the pixel values for the bounding boxes.For example, if the image size is 1000x1000 pixels and the bounding box value is [0.6822429906542056,0.07027328499721137], you can calculate the pixel values as follows:
x1 = 0.6822429906542056 * 1000 = 682.2429906542056
y1 = 0.07027328499721137 * 1000 = 70.27328499721137
To use this formula, you need to know the size of the image in pixels and the normalized coordinates for the bounding boxes from the labels.json file.
You can then use the formula to calculate the pixel values for the bounding boxes. Hope this helps.
If this answers your query, do click Accept Answer
and Yes
for was this answer helpful.