删除 scikit-image 中的标签 [英] Removing labels in scikit-image

问题描述

我已经标记了一个二值图像

I have labeled an binary image

imageLabels = morphology.label(imageBinary, background=255)

但是，当我检查标签数量时，我得到了 535 个元素.

However when I check the number of labels, I get 535 elements.

print(len(imageLabels))

作为一个解决方案，我考虑使用 measure.regionprops 来移除具有小像素区域的标签.你们会如何处理这个问题?我尝试了以下方法，但出于某种原因，新数组不再被视为正确的标签元素.

As a solution for this I thought about using measure.regionprops in order to remove the labels with a small pixel area. How would you guys approach this? I have tried the following, but for one reason or another the new array is no longer seen as a correct label element.

i=0
for labelprop in measure.regionprops(imageLabels):
    if (labelprop.area > 100):
        imageLabels_keep.append(imageLabels[i])
    i=i+1

推荐答案

我认为 morphology.remove_small_objects(image, min_px_size) 可以满足您的需求.下面是一个使用该函数的示例:

I think morphology.remove_small_objects(image, min_px_size) does what you're looking for. Here's an example that uses that function:

http://scikit-image.org/docs/dev/auto_examples/applications/plot_coins_segmentation.html#edge-based-segmentation

这篇关于删除 scikit-image 中的标签的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！