Lung Segmentation

 Lung Segmentation
First, for use of the 3D image-processing and analysis
technique, we used a trilinear interpolation technique to
make the size of the voxels equal to 1 mm in each of
three dimensions. We then separated lung regions from
other regions and background outside the body for each
of the CT sections by use of a thresholding technique. A
pixel with a CT value between 400 HU and 1000 HU
was considered to be located inside the lung region and
was thus assigned a value of 1; otherwise, the pixel was
assigned a value of 0. After the thresholding, the lung
regions, together with background (air) regions outside
the body, were separated from other regions inside the
body. In each section, we removed air regions outside the
body by discarding regions of value 1 that were attached
to image boundaries. The regions that survived this processing
were considered to be lung regions.
However, nodules connected to the pleura might have
been excluded from the lung regions since the CT values
for the pixels inside the nodule were outside the range of
400 HU to 1000 HU. To correct this type of segmentation
error, we first tracked the contour of the lung region,
shown as a solid curve in Figure 3. From a starting
contour point, we then scanned counterclockwise all
points on the entire contour one by one. For each current
contour point “A,” we then scanned clockwise all contour
points between the starting point and point “A” until we
found a point “B” that satisfied the following conditions:
(1) the distance between points “A” and “B” is less than
30 mm, (2) all the pixels on a straight line “AB” connecting
points “A” and “B,” shown as a dotted line in Figure 3,
have a value of 0 in the lung-segmented image, and (3)
the maximum distance between the straight line “AB” and
all the contour points between points “A” and “B” on the
contour is larger than two thirds of the distance between
points “A” and “B.” When such a point “B” was found,
we assumed that we had found a juxtapleural object and
we replaced all the contour points between points “A”
and “B” by the dotted straight line “AB.” By doing so,
we included a juxtapleural object inside the lung region.
If such a point “B” could not be found, we simply proceeded
to the next contour point of “A” and considered it
as the new current point. From the new current point, we
repeated the above procedure until all contour points were
checked.

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
The human visual cortex is biased towards shape components while CNNs produce texture biased features. This fact may explain why the performance of CNN significantly degrades with low-labeled input data scenarios. In this paper, we propose a frequency re-calibration U-Net (FRCU-Net) for medical image segmentation. Representing an object in terms of frequency may reduce the effect of texture bias, resulting in better generalization for a low data regime. To do so, we apply the Laplacian pyramid in the bottleneck layer of the U-shaped structure. The Laplacian pyramid represents the object proposal in different frequency domains, where the high frequencies are responsible for the texture information and lower frequencies might be related to the shape. Adaptively re-calibrating these frequency representations can produce a more discriminative representation for describing the object of interest. To this end, we first propose to use a channel-wise attention mechanism to capture the relationship between the channels of a set of feature maps in one layer of the frequency pyramid. Second, the extracted features of each level of the pyramid are then combined through a non-linear function based on their impact on the final segmentation output. The proposed FRCU-Net is evaluated on five datasets ISIC 2017, ISIC 2018, the PH2, lung segmentation, and SegPC 2021 challenge datasets and compared to existing alternatives, achieving state-of-the-art results.请详细介绍这段话中的技术点和实现方式
最新发布
05-29
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值