[英]OpenCV - haarcascade training
is it normal, that at stage 6 I have results like this?这正常吗,在第6阶段我有这样的结果? Sample percentage is around 0-3% at 1000+ features and FA is around 0.8.
超过 1000 个特征的样本百分比约为 0-3%,FA 约为 0.8。 Would love to get some insight, also the stage 5 has only 50 used features while stage six has 1500+ and it keeps growing.
很想了解一下,第 5 阶段只有 50 个使用的功能,而第 6 阶段有 1500 多个,并且还在不断增长。
Setting is like this, Iam training cascade to recognize faces.设置是这样的,我训练级联识别人脸。
The answer is, that I anotated my ROIs poorly and it was learning basically everything that was on the dataset pictures.答案是,我对 ROI 的注释很差,它基本上学习了数据集图片上的所有内容。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.