如何将label-studio的rle格式转换为黑白图像蒙版答案

【问题标题】：how to convert rle format of label-studio to black and white image masks如何将label-studio的rle格式转换为黑白图像蒙版
【发布时间】：2022-12-24 00:24:39
【问题描述】：

给定值数组如下：完整数组here

[{"format": "rle", "rle": [0, 18, 192, 0, 57, 27, 255, 255, 255, 0, 259, 96, 17, 192, 97, 255, 248, 239, 227, 16, 2....., 255, 142, 3, 130, 21, 128, 0], "brushlabels": ["crack"], "original_width": 640, "original_height": 480}]

我怎样才能转换成 image.png？

【问题讨论】：

标签： python opencv image-processing

【解决方案1】：

您可以使用 np.reshape() 方法将数组重塑为图像的原始尺寸：

import cv2
import numpy as np

data = [{"format": "rle", "rle": [255, 0, 0, 0, 0, 0, 0, 0, 255, 0, 0, 255, 0, 0, 0, 255, 0, 0], "brushlabels": ["crack"], "original_width": 2, "original_height": 3}]

# Create image from data
img = np.uint8(data[0]["rle"]).reshape(data[0]["original_height"], data[0]["original_width"], 3)

# The image is completed, but it's tiny small in this example, so resize it to be able to visualize it
img = cv2.resize(img, (200, 300), interpolation=cv2.INTER_NEAREST)

# Show result
cv2.imshow("Result", img)
cv2.waitKey(0)

输出：

【讨论】：

您好，感谢您的回复。 rle 以不存储 0 的方式对数据进行编码。 rle 数据的形状是(8160,)，而图像是(480x640)。简单的重塑在这里不起作用。我在 github gist gist.github.com/aliwaqas333/eb87c88ea0edb157e046e97350ddd9b6 中附上了完整的变量

【解决方案2】：

基于https://github.com/heartexlabs/label-studio-converter/blob/master/label_studio_converter/brush.py#L70

from typing import List
import numpy as np


class InputStream:
    def __init__(self, data):
        self.data = data
        self.i = 0

    def read(self, size):
        out = self.data[self.i:self.i + size]
        self.i += size
        return int(out, 2)


def access_bit(data, num):
    """ from bytes array to bits by num position"""
    base = int(num // 8)
    shift = 7 - int(num % 8)
    return (data[base] & (1 << shift)) >> shift


def bytes2bit(data):
    """ get bit string from bytes data"""
    return ''.join([str(access_bit(data, i)) for i in range(len(data) * 8)])


def rle_to_mask(rle: List[int], height: int, width: int) -> np.array:
    """
    Converts rle to image mask
    Args:
        rle: your long rle
        height: original_height
        width: original_width

    Returns: np.array
    """

    rle_input = InputStream(bytes2bit(rle))

    num = rle_input.read(32)
    word_size = rle_input.read(5) + 1
    rle_sizes = [rle_input.read(4) + 1 for _ in range(4)]
    # print('RLE params:', num, 'values,', word_size, 'word_size,', rle_sizes, 'rle_sizes')

    i = 0
    out = np.zeros(num, dtype=np.uint8)
    while i < num:
        x = rle_input.read(1)
        j = i + 1 + rle_input.read(rle_sizes[rle_input.read(2)])
        if x:
            val = rle_input.read(word_size)
            out[i:j] = val
            i = j
        else:
            while i < j:
                val = rle_input.read(word_size)
                out[i] = val
                i += 1

    image = np.reshape(out, [height, width, 4])[:, :, 3]
    return image

用法：

image = rle_to_mask(
    result['value']['rle'], 
    result['original_height'], 
    result['original_width']
)
print(image.shape)  # (original_height, original_width)

from PIL import Image
Image.fromarray(image).show()

【讨论】：