mikucy / cs131 Goto Github PK

in hw8, at iterative_lucas_kanade function
you write:
Ik = img1[y1, x1] - img2[y2, x2]
bk = np.array([np.sum(IkA1), np.sum(IkA2)])
But in equations, Ik is an array which has different values at (x,y) with x = px-w-->px+w, y=py-w-->py+w, not just a number (your code)
In my opinion, it would be Ik = img1[y1-w:y1+w+1,x1-w:x1+w+1] - img2[y2-w:y2+w+1,x2-w:x2+w+1]
with y2 = int(round(y+gy+vy)); x2 = int(round(x+gx+vx)) as mentioned in the code

   Hello, I notice that you didn't complete the remove_object in the last of the hw4. I complete it, maybe not perfect.
  Below is my code:

def remove_object(image, mask):
    """Remove the object present in the mask.

    Returns an output image with same shape as the input image, but without the object in the mask.

    Args:
        image: numpy array of shape (H, W, 3)
        mask: numpy boolean array of shape (H, W)

    Returns:
        out: numpy array of shape (H, W, 3)
    """
    out = np.copy(image)

    ### YOUR CODE HERE
    from skimage import measure
    label_image = measure.label(mask)
    regions = measure.regionprops(label_image)
    region = regions[0]
    if len(regions) != 1:
        print("Maybe two objects to remove?")
        # Find the biggest area of region
        for i in regions:
            if i.area > region.area:
                region = i
    transposeImage = False
    if region.bbox[2] - region.bbox[0] < region.bbox[3] - region.bbox[1]:
        out = np.transpose(out, (1, 0, 2))
        mask = np.transpose(mask, (1, 0))
        transposeImage = True
    count = 0   # count time for all iteration
    while not np.all(mask == 0):
        energy_image = energy_function(out)
        energy_image = energy_image + energy_image * mask * (-1000)
        vcost, vpaths = compute_forward_cost(out, energy_image)
        end = np.argmin(vcost[-1])
        seam = backtrack_seam(vpaths, end)
        out = remove_seam(out, seam)
        mask = remove_seam(mask, seam)
        count += 1
    #print("count = ", count)
    out = enlarge(out, out.shape[1] + count)
    if transposeImage:
        out = np.transpose(out, (1, 0, 2))
    ### END YOUR CODE
    return out

The result is:

I notice that the result is not perfect, below is the picture provided by.
http://cs.brown.edu/courses/cs129/results/proj3/taox/

In the top right have some subtle change. I don't know how to do better. If you have some ideas
to improve it, we could achieve best result.

Hello, I am a beginner. In hw3_release Extra Credit: Stitching Multiple Images, I don't know how to do.
The problem said given a sequence of m images, then stitch an ordered chain of images. I want to take the first and second images then stitch together, formed a new images, then stitch together the third images. But it is too complex and maybe have some simple ways.
I notice that you haven't do this problem, if you have any idea, please complete the code. I am a beginner, it is too hard for me. Thank you so much.

Hello,
When I am working with the function find_seams, I have got the error like this:

IndexError Traceback (most recent call last)
in ()
3 # Alternatively, find k seams for removal and duplicate them.
4 start = time()
----> 5 seams = find_seams(img, W_new - W)
6 end = time()
7

~/Desktop/cv&nlp/CS131_release/hw4_release/seam_carving.py in find_seams(image, k, axis, efunc, cfunc)
331 # Store the new seam with value i+1 in the image
332 # We can assert here that we are only writing on zeros (not overwriting existing seams)
--> 333 assert np.all(seams[np.arange(H), indices[np.arange(H), seam]] == 0), "we are overwriting seams"
334 seams[np.arange(H), indices[np.arange(H), seam]] = i + 1
335

IndexError: arrays used as indices must be of integer (or boolean) type

but I didn't modify the code. Have you got any problems like this? How can I fix this error?
Hope for your reply, thanks a lot!