Ok - this guy has already built an app for real-time object recognition, see here: https://github.com/datitran/object_detector_app
You can play around with it for fun!
Edit: If in case anyone was wondering why the window can’t close properly… I suspect you need to press ‘q’ to exit.
However, as my purpose is quite different, I shall attempt to build my own video detection thingy based on the Tensorflow API (the frozen model first, then we shall see if I want to unfreeze the top layers and train it) LOL
Not going to be easy! First things first… how do I get my video into frames?
import numpy as np import cv2 cap = cv2.VideoCapture('alligator_video.mp4') # you can change it to your video file while(cap.isOpened()): ret, frame = cap.read() if ret == True: cv2.imshow('frame', frame) if cv2.waitKey(1) & 0xFF == ord('q'): break else: break cap.release() cv2.destroyAllWindows()
I’m just trying to get it to show each frame of the video for now… but somehow my video file can’t close at cap.release() which is strange. Guess I have to slowly learn from scratch and figure it out.
OpenCV, here I come!
Edit: This might be more helpful for general use: http://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_setup/py_intro/py_intro.html#intro
On a side note, if anyone has any trouble installing opencv on python 3 (mine kept installing on python 2.7 and pip3 didn’t work), here’s what I used (thanks to my trusty personal advisor):
python3 -m pip install opencv-python
And if anyone is ever curious what
-m means: https://docs.python.org/3/using/cmdline.html