yashmukaty / harry-potters-invisible-cloak

Image masking basic techniques

Harry-Potters-Invisible-cloak

Image masking basic techniques

Tired of scratching your head through hectic reasearch papers? Common let's make something interesting with the help of basic computer vision knowledge using opencv library.

Lets run the code first:

• Install all dependencies:

pip install -r requirements.txt

• Run code:

python src.py

Stay away from frame untill video window pops up for the background image to be created

What's in the code:

The code has been written to recognize RED color as cloak for now.

lower_range = np.array([0,120,50])
higher_range = np.array([10,255,255])

You can change the color bound according to yourself.The color range for HSV value are available here. For HSV, Hue range is [0,179], Saturation range is [0,255] and Value range is [0,255]. Different softwares use different scales. So if you are comparing OpenCV values with them, you need to normalize these ranges.

• Masking is one of the key features for this project:

mask1 = cv2.inRange(hsv, lower_range, higher_range)

The above forms a mask of the area we want to make invisible to frame-feed.

• The next task is to extract the above mask from the frame,background and foreground.

#segmenting out cloth color
mask2 = cv2.bitwise_not(mask1)

#Segment the red color part out of the frame using bitwise and with the inverted mask
layer1 = cv2.bitwise_and(img, img, mask = mask2)

#Create image showing static background frame pixels only for the masked region
layer2 = cv2.bitwise_and(background, background, mask = mask1)

• This is now where the magic happens:

#output
output = cv2.addWeighted(layer1,1,layer2,1,0)

Both the mask were added with equal weights using opencv's addWeight

References:

www.machinelearningman.com

https://docs.opencv.org/master/d0/d86/tutorial_py_image_arithmetics.html