仅使用 imageai.Detection 检测一个人

Question

I am new to imageAI and recently I read an article about a small project of detecting pedestrians. I try to get the same result, but some problems occurred.

Here is my code:

from imageai.Detection import ObjectDetection
import os

execution_path = os.getcwd()

detector = ObjectDetection()
detector.setModelTypeAsRetinaNet()
detector.setModelPath(os.path.join(execution_path , "resnet50_coco_best_v2.1.0.h5"))
detector.loadModel()
custom_objects = detector.CustomObjects(person=True, car=False)
detections = detector.detectCustomObjectsFromImage(input_image=os.path.join(execution_path , "image.png"), output_image_path=os.path.join(execution_path , "image_new.png"), custom_objects=custom_objects, minimum_percentage_probability=65)


for eachObject in detections:
    print(str(eachObject["name"]) + " : " + str(eachObject["percentage_probability"]))
    print("--------------------------------")

# show the image
from IPython.display import Image
Image("image_new.png")

There's a warning and also the result printed:

WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.

<ipython-input-11-277100bcf064>:11: DeprecationWarning: 'detectCustomObjectsFromImage()' function has been deprecated and will be removed in future versions of ImageAI. 
 Kindly use 'detectObjectsFromImage()' 
  detections = detector.detectCustomObjectsFromImage(input_image=os.path.join(execution_path , "image.png"), output_image_path=os.path.join(execution_path , "image_new.png"), custom_objects=custom_objects, minimum_percentage_probability=65)

WARNING:tensorflow:6 out of the last 6 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbfc2d4df70> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.
person : 72.94188737869263
--------------------------------

Here is the photo I input: original And here is the resulting photo: new

The original photo has no box, and the new one has only one box. But the new one should have two boxes according to two persons.

I don't know if only one box is due to the warning or something else? Could someone help me out?

Thanks.

Answer 1

This is the code i use, i think ImageAI uses the box_points while running to differentiate between multiple objects, could be very wrong tho.

for eachObject, eachObjectPath in zip(detections, objects_path):
    print(eachObject["name"] , " : " , eachObject["percentage_probability"], " : ", eachObject["box_points"] )
    print("Object's image saved in " + eachObjectPath)
    print("--------------------------------")