机器学习在图像检测方面主要有 分类、目标检测、分割 三个任务
分类——CNN
只是把图片做classify
LeNet,ResLet
目标检测——SSD
更进一步,在图像上标出object的位置,并且object可能不止一个
Understanding SSD MultiBox — Real-Time Object Detection In Deep Learning
“A few years ago, by exploiting some of the leaps made possible in computer vision via CNNs, researchers developed R-CNNs to deal with the tasks of object detection, localization and classification.
Fortunately, in the last few years, new architectures were created to address the bottlenecks of R-CNN and its successors, enabling real-time object detection. The most famous ones are YOLO (You Only Look Once) and SSD MultiBox (Single Shot Detector). ”
UNet——分割
像素级别的识别
FCN,UNet
R-CNN Fast-R-CNN Faster-R-CNN Mask-Faster-R-CNN
RCNN (论文:Rich feature hierarchies for accurate object detection and semantic segmentation) 是将CNN方法引入目标检测领域, 大大提高了目标检测效果,可以说改变了目标检测领域的主要研究思路, 紧随其后的系列文章:( RCNN),Fast RCNN, Faster RCNN 代表该领域当前最高水准。
这个领域现在可是变得太快了,过一两年就把前面的给颠覆了。