utils.py

import numpy as np


EPS = 1e-10


def dice(true, pred):
    true = true.astype(bool)
    pred = pred.astype(bool)

    intersection = (true & pred).sum()
    im_sum = true.sum() + pred.sum()

    return 2.0 * intersection / (im_sum + EPS)


def dice_all(true, pred):
    return np.mean([dice(t, p) for t, p in zip(true, pred)])


def rle_encode(img):
    '''
    img: numpy array, 1 - mask, 0 - background
    Returns run length as string formated

    from: https://www.kaggle.com/kmader/baseline-u-net-model-part-1
    '''
    pixels = img.flatten()
    pixels = np.concatenate([[0], pixels, [0]])
    runs = np.where(pixels[1:] != pixels[:-1])[0] + 1
    runs[1::2] -= runs[::2]
    return ' '.join(str(x) for x in runs)


def rle_decode(mask_rle, shape=(320, 240)):
    '''
    mask_rle: run-length as string formated (start length)
    shape: (height,width) of array to return
    Returns numpy array, 1 - mask, 0 - background

    from: https://www.kaggle.com/kmader/baseline-u-net-model-part-1
    '''
    s = mask_rle.split()
    starts, lengths = [np.asarray(x, dtype=int) for x in (s[0:][::2], s[1:][::2])]
    starts -= 1
    ends = starts + lengths
    img = np.zeros(shape[0]*shape[1], dtype=np.uint8)
    for lo, hi in zip(starts, ends):
        img[lo:hi] = 1

    return img.reshape(shape).T