post processing implementation of detector ONNX javascript implementation

[pashanitw]

Hi this is a great repo thanks for open sourcing. i exported onnx model and trying to run detector in javascript. but seems der are lot of post processing involved for the outputs of onnx output. has anyone implemented how to obtain the boxes from the onnx outputs? or please shed some light on how to implement in javascript. thanks

[MdotO]

Not sure if this is useful but you can use this code in python : https://github.com/JaidedAI/EasyOCR/blob/master/easyocr/DBNet/DBNet.py You need to convert hmap2bbox python function to equivalent js function to get the bounding boxes