Creating interfaces

Vuba provides a series of OpenCV HighGUI contructors for easily creating image analysis user interfaces. All of these contructors are based around the base constructor class: BaseGUI. This class provides a series of decorators that allow us to code up complex interfaces using simple, declarative code.

Single trackbar interfaces

To demonstrate the ease of use of this module, let’s first code up a simple binary threshold viewer containing a single trackbar. We will first need to grab some footage to work with:

import vuba
import cv2

# Create a handler for reading in the video
video = vuba.Video("../examples/example_data/raw_video/test.avi")

# Read in the first frame for the viewer
first = video.read(index=0)

Now that we have some footage, we can go about creating our user interface. First, we will need to initiate an instance of FrameGUI since we are working with a single frame:

# Initiate an instance of FrameGUI
gui = vuba.FrameGUI(first, "Threshold viewer")

Next we need to integrate our image processing function into the interface. Since we are only doing a binary threshold here this is relatively simple:

@gui.method
def threshold(gui):
    frame = gui.frame.copy()  # create a copy of the current frame
    thresh_val = gui["thresh_val"]  # grab the current threshold value

    # Grayscale the current frame
    gray = vuba.gray(frame)

    # Threshold the current frame
    _, thresh = cv2.threshold(gray, thresh_val, 255, cv2.THRESH_BINARY)

    return thresh

Finally, we need to add a trackbar so that we can visualise the effects of varying our binary threshold:

@gui.trackbar("Threshold", id="thresh_val", min=0, max=255)
def on_thresh(gui, val):
    gui["thresh_val"] = val
    img = gui.process()
    cv2.imshow(gui.title, img)

We’re almost done, now all we need to do is run the gui we created:

gui.run()

The output of this code should give you the following gui:

This might seem trivial, but in less than 20 lines of code (excluding comments and whitespace) we have created a user interface that enables us to view the output of a variable binary threhsold. Note that because all gui constructors use the same base class, we only have to address the video constructor and supply an object that describes a video instead, to apply the interface to a movie:

# Initiate an instance of VideoGUI
gui = vuba.VideoGUI(video, "Threshold viewer")

@gui.method
def threshold(gui):
    frame = gui.frame.copy()  # create a copy of the current frame
    thresh_val = gui["thresh_val"]  # grab the current threshold value

    # Grayscale the current frame
    gray = vuba.gray(frame)

    # Threshold the current frame
    _, thresh = cv2.threshold(gray, thresh_val, 255, cv2.THRESH_BINARY)
    return thresh

@gui.trackbar("Threshold", id="thresh_val", min=0, max=255)
def on_thresh(gui, val):
    gui["thresh_val"] = val
    img = gui.process()
    cv2.imshow(gui.title, img)

# And finally execute the gui
gui.run()

Multi-trackbar interfaces

To add additional trackbars to an interface, we only need to specify another trackbar using BaseGUI.trackbar and add additional steps to our image processing method:

# Initiate an instance of VideoGUI
gui = vuba.VideoGUI(video, "Threshold viewer")

@gui.method
def threshold(gui):
    frame = gui.frame.copy()  # create a copy of the current frame
    tb_vals = gui.values() # grab all current trackbar values

    # Grayscale the current frame
    gray = vuba.gray(frame)

    # Threshold the current frame
    _, thresh = cv2.threshold(gray, tb_vals['thresh_val'], 255, cv2.THRESH_BINARY)

    # Blur the thresholded image (note to exclude even values for actual applications)
    blur = cv2.GaussianBlur(thresh, (tb_vals['blur_val'],tb_vals['blur_val']), 0, 0)

    return blur

# Add trackbars using the default callback
gui.trackbar("Threshold", id="thresh_val", min=0, max=255)(None)
gui.trackbar("Blur", id="blur_val", min=1, max=101)(None)


# And finally execute the gui
gui.run()

This should give us the following interface:

Creating interfaces

Single trackbar interfaces

Multi-trackbar interfaces

See also