Question 1: Part 1: i) Run this command python Q11.py ii) It will create input folder in the same directory and save images when any face is detected until Space key is pressed. iii) Now after space key if Esc is pressed,it will start saving images in input folder whenever face is detected. iv) It will stop taking images if 's' key is pressed. Note: 1) This execution requires internet as it will download .bat files if any required file is missing. 2) This file need align.py file in the same directory.
Part 2: i) Run this command python Q1.py -p shape_predictor_68_face_landmarks.dat ii) It will calculate all the keypoints coordinates and save it into two .npy files which are image_train.npy and landmark.npy. iii) image_train.npy contains all the images and landmark.npy contains 4 facial keypoints. iv) This 4 points are x and y coordinates of chin,left eye,right eye and forehead in landmark.npy file.
Part 3: i) Run every cell of Face_landmark.ipynb ii) It will show you all the graph and one prediction at last.
Question 2: i) Run this command python Q2.py ii) Give video file path as input if video is present in the same directory just give object_tracking_example.mp4 as input. iii) I have generalize this question so whenever you press 'p' it will point the green ball(object) by an arrow for five second. iv) You can quit by pressing 'q'. Note : Download video file from this link https://www.dropbox.com/s/mykm9l29uz155tr/object_tracking_example.mp4?dl=1
Important Links:
- Link for input folder: https://drive.google.com/open?id=1xoVd07qw1yZFgOYI2IPPA6rosbkTIM69
- Link for input_data folder: https://drive.google.com/open?id=1BAShvf8axgai2T5sCahq7ZTNBOuT_pw4
- Link for shape_predictor_68_face_landmarks.dat file: https://drive.google.com/open?id=1RiHjIumKhQGAXW7EA8qNd1A9FtoNcYm1
- Link for image_train.npy : https://drive.google.com/open?id=1Yffp_zFq9k54K2tAe8SWpfbatxD5P1lg
- Link for models folder: https://drive.google.com/open?id=1K4PxY0phf21_JIl2EPmc-T7iYyNvOpY7