Improla == IMage PROcessing LAb. It is meant for image processing research and teaching. From ImprolaLib-0.0.2, it supports reading and writing PNG images. It uses PNGwriter(pngwriter.sourceforge.net) for this purpose.-Improla == IMage PROcessing LAb

that sourc help u to find sum shape in picture using emgu cv warrper of open cv for .net project

使用Pytorch搭建U-Net，该模型可以对随机传入任意大小的图片进行图片分割，根据所训练的数据和标签得到索要分割的区域。(Using Python to build u-net, the model can segment random incoming pictures of any size, and get the region to be segmented according to the trained data and labels.)

UNet最早发表在2015的MICCAI上，短短3年，引用量目前已经达到了4070，足以见得其影响力。而后成为大多做医疗影像语义分割任务的baseline，也启发了大量研究者去思考U型语义分割网络。而如今在自然影像理解方面，也有越来越多的语义分割和目标检测SOTA模型开始关注和使用U型结构，比如语义分割Discriminative Feature Network(DFN)(CVPR2018)，目标检测Feature Pyramid Networks for Object Detection(FP