Xuebin Qin

There are mainly three reasons for triggering Illegal instruction(core dumped): (1) Libraries missing. (2) bad memory operation, e.g. read a deleted vector or array, (3) "M ay be you compiled...

Hi, lineSylClr[i],linewidth=linewidth[i] have to be manually designed in the line 21 and 22: lineSylClr = ['r-','b-'] # curve style, same size with rs_dirs linewidth = [2,1] # line width, same...

please check if your data is with the right channels as well as the the range [0,1.0] or [0,255]. On Fri, Oct 21, 2022 at 4:03 AM 锦时 ***@***.***> wrote:...

Thanks for your interest. The is mainly because of the version issues. You can try the following ways: （1）install Ubuntu 14.04 or 16.04 64bit， （2） if you don't want to...

有如下两种方法可以尝试： （1）安转一个Ubuntu 14.04或者16.04 64bit的虚拟机，在里面跑试一下， （2） 你也可以尝试自己编译以下，首先将Makefile的line 17改成：$(lnk) $(dbg) -o bylabel $(objRRC) `pkg-config --libs opencv` EDLib.a EDLinesLib.a （因为我们之前的Makefile用的是静态编译，我们把所有依赖库都加进去了，但是其实如果你重新动态编译，这些库都是不需要的），然后make一下试试，但是在make编译之前需要安装opencv 2.4.9或者之后的版本，而且在安装的时候需要，保持QT=enable,否则无法通过鼠标滚轮缩放图片。

To achieve results like remove.bg, larger and wider datasets have to be used and more engineering works have to be done. We can't expect that a model trained on a...

Yes, true. The supervisely dataset is more accurate than COCO but still misses some structures and more details. Glad to hear that you are creating new datasets. Yes, dataset contributes...

1. image with RGB channels [0,255], Mask with one channel (0-background, 255-foreground). 2. yes, you can train with 'train.py' directly. just modify the training data path should work. If you...

Sorry for my mistake. We just uploaded the EDLib.a. Thanks for your correction. On Wed, Feb 26, 2020 at 3:53 PM cctt126 wrote: > Thanks for sharing the code. Can...

the input resolution is highly influencing the receptive field so that lead to different "data distribution" when you change the resolution. The accuracy degradation kind of make sense under current...