Frame selection, resizing, and grayscale conversion

• We want to resize each frame to be 720 pixels in width and 480 pixels in height.
  • Set target_size to (720, 480)
• All the frames are not required. Let’s just capture exactly 12 random frames from the video.
  • Set mode to "random"
• And finally, visualize the captured frames using mydia.make_grid()

# Imports
import matplotlib.pyplot as plt
from mydia import Videos, make_grid

# Initialize video path
video_path = r"./sample_video/bigbuckbunny.mp4"

# Configuring the parameters
# For other paramaters available, view the code documentation.
reader = Videos(target_size=(720, 480), num_frames=12, mode="random")

# Call the 'read()' function to get the required video tensor
# which will be of shape (1, 12, 480, 720, 3)
video = reader.read(video_path)
print("The shape of the tensor:", video.shape)

# Plot the video frames in a grid
grid = make_grid(video[0])
plt.imshow(grid)

Out:

The shape of the tensor: (1, 12, 480, 720, 3)

Note

The number of channels for a RGB video is 3 (indicated by the last value in the tuple).

• Now let’s read the video with the same configuration, but in grayscale
  • For this, set to_gray to True
• Also, the function make_grid() takes certain arguments to construct the grid of frames of the video. For more info, view mydia.make_grid().

# Imports
import matplotlib.pyplot as plt
from mydia import Videos, make_grid

# Initialize video path
video_path = r"./sample_video/bigbuckbunny.mp4"

# Configuring the parameters
reader = Videos(target_size=(720, 480), to_gray=True, num_frames=12, mode="random")

# Call the 'read()' function to get the required video tensor
# which will be of shape (1, 12, 480, 720, 1)
video = reader.read(video_path)
print("The shape of the tensor:", video.shape)

# Plot the video frames in a grid
grid = make_grid(video[0], num_col=2)
plt.imshow(grid, cmap="gray")

Out:

The shape of the tensor: (1, 12, 480, 720, 1)

Note

The number of channels for a video in gray scale is 1 (indicated by the last value in the tuple).

Total running time of the script: ( 0 minutes 1.137 seconds)