Python Deep Learning Projects

Encoding

With the make_normalize() function defined, let's now define a make_target function. This function accepts the segmentation array of shape (360,480) and then returns a segmentation target of shape (360,480,2). In the target, channel 0 represents the background and will have 1 in locations that represent the background in the image and zero elsewhere. Channel 1 represents the person and will have 1 in locations that represent the person in the image and 0 elsewhere. The following code implements the function:

def make_target(labels):    """Function to one hot encode targets."""    x = np.zeros([360,480,2])    for i in range(360):        for j in range(480):            x[i,j,labels[i][j]]=1    return xplt.figure(figsize = (14,5))plt.subplot(1,2,1)plt.imshow(make_target(lab_lst[0])[:,:,0]) ...