From 635ca681e1c516e5a5e0b7af2e826c170adb070e Mon Sep 17 00:00:00 2001 From: zeek <984294471@qq.com> Date: Sat, 25 Apr 2020 17:09:18 +0800 Subject: [PATCH] =?UTF-8?q?=E8=B7=9F=E6=96=B0=E7=9B=AE=E6=A0=87=E6=A3=80?= =?UTF-8?q?=E6=B5=8B=E5=9B=BE=E7=89=87=E9=93=BE=E6=8E=A5?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- targetDetection/README.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/targetDetection/README.md b/targetDetection/README.md index 569bbff..82bc210 100644 --- a/targetDetection/README.md +++ b/targetDetection/README.md @@ -7,10 +7,12 @@ 3. 检测-Detection:解决“是什么?在哪里?”的问题,即定位出这个目标的的位置并且知道目标物是什么。 4. 分割-Segmentation:分为实例的分割(Instance-level)和场景分割(Scene-level),解决“每一个像素属于哪个目标物或场景”的问题。 -![pic](http://index.zeekling.cn/gogsPics/ml/targetDection/1.png) +![pic](https://img.zeekling.cn/images/2020/04/25/63e746f0962e6215b5ed6a2587fe5af0.png) ## 目标检测算法分类 -![pic](http://index.zeekling.cn/gogsPics/ml/targetDection/2.png) + +![pic](https://img.zeekling.cn/images/2020/04/25/edfbb1a6466964df97d6bea080496c67.png) + ### Two stage目标检测算法 先进行区域生成(region proposal,RP)(一个有可能包含待检物体的预选框),再通过卷积神经网络进行样本分类。