Replication Data for: 2017-12-06: DIP/WT Images (8hrs x5 magnification) - Analysis Results (new settings)

Replication Data for: 2017-12-06: DIP/WT Images (8hrs x5 magnification) - Analysis Results (new settings): settings and new parameters: 4PX and 15PX We tried to resolve an issue with the "large circles" in images. It required adding a new parameter - "Relative intensity variation in a spot". I think that its adequate value is 0.25 or 0.5. Also please note that currently there are two parameters that control minimal spot size - "Minimal linear size of an object" and "Min number of pixels in one object". Probably the second parameter can be excluded from the interface as a control solely via linear size (radius) seems to be sufficient. Anyway, for the 5x images acceptable min linear size can be set to 10 (or may be even to a lower value to separate small closely located spots), and min number of pixels can be set to 50-100 (or again even smaller - e.g., even min radius 4 with min number of pixels 15 seem to be ok for a subset of images we used for testing). (Generally if one does not want to think about a value of min number of pixels, he can set it, e.g., to a square of min_radius or a half of this square).

复现数据：2017年12月6日：DIP/WT图像（8小时×5倍放大）—— 新设置下的分析结果：设置与新增参数：4PX与15PX。本次工作旨在解决图像中"大圆圈"相关问题，为此新增了一项参数——"斑点相对强度变化量（Relative intensity variation in a spot）"。笔者认为其合理取值为0.25或0.5。另需说明：当前共有两项参数用于控制最小斑点尺寸，分别为"目标最小线性尺寸（Minimal linear size of an object）"与"单目标最小像素数（Min number of pixels in one object）"。鉴于仅通过线性尺寸（半径）即可实现有效控制，第二项参数或可从操作界面中移除。无论如何，针对5倍放大图像，可接受的最小线性尺寸可设为10（甚至可调低以分离间距较近的小型斑点），而单目标最小像素数可设为50至100（同样可进一步调低——例如在我们的测试子集图像中，最小半径设为4、最小像素数设为15即可取得良好效果）。（一般而言，若无需手动调整最小像素数参数，可将其设为最小半径的平方，或该平方值的一半即可。）