How to the get confidence score of object?

I have referenced the article below.

Continuing the discussion from How to Display confidence with label in deepstream ? like (person 0.81)

I want show the confidence to sreen. Guide me.

1 Like

The object confidence is stored in parameter obj_meta->confidence, you can set it in obj_meta->text_params.display_text and it can be displayed on the bounding box.
You may refer to the relevant topic here: Show detection confidente along with the class label

I used Python. Do you have any other suggestions for me?

Still using deepstream_test_1.py as example, add last line as below can display confidence. You can customize obj_meta.text_params.display_text to any string if you want.

        l_obj=frame_meta.obj_meta_list
        while l_obj is not None:
            try:
                # Casting l_obj.data to pyds.NvDsObjectMeta
                obj_meta=pyds.NvDsObjectMeta.cast(l_obj.data)
            except StopIteration:
                break
            obj_counter[obj_meta.class_id] += 1
            obj_meta.rect_params.border_color.set(0.0, 0.0, 1.0, 0.8) #0.8 is alpha (opacity)
            print("object id={}, confidence={}\n".format(obj_meta.class_id, obj_meta.confidence))
            obj_meta.text_params.display_text = "%4f"%obj_meta.confidence

I’m grateful for your help!

I have one more problem that I can’t solve, class_id is 0,1,2,… How to convert obj.class_id to class_name when screen?

id and name are bound one by one, You can refer to the code below

py_nvosd_text_params.display_text = “Frame Number={} Number of Objects={} Vehicle_count={} Person_count={}”.format(frame_number, num_rects, obj_counter[PGIE_CLASS_ID_VEHICLE], obj_counter[PGIE_CLASS_ID_PERSON])

1 Like

define an array for class name and reference it via class_id:

class_names = [ 'vehicle', 'bicycle','person', 'roadsign']

...


obj_meta.text_params.display_text = "%s:%4f"%(class_names[obj_meta.class_id], obj_meta.confidence)
1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.