Hello Sir/Madame,
I have checked the solution and I can tell why is it too slow.
I can recommend a different and faster way.
Here is a little summary about me:
I have an extensive experience in computer vision and deep learning applications with a great knowledge of python, numpy, matplotlib, in addition to image processing libraries like OpenCV and PIL.
I am also an expert in working with machine learning libraries like TensorFlow and Pytorch.
I have dealt with state of art object detection and classification models like VGG16, MobileNetv2, YOLOV3, SSD, among others.
One of my most recent projects is where I used YOLOV3 for a traffic signs detection and recognition system and utilized VGG16 to design an in campus localization application. Moreover, I did a research for compressing the YOLOV3 model and produced a lightweight architecture which is 95% smaller than the original and the research was translated into a conference publication. I am currently focusing on producing a deep learning model to enhance images taken under low light conditions and looking forward for producing a very efficient design.
I hope we can chat and discuss further details regarding the project.
Looking forward to hearing from you!