# Detection of Emotions ## Detection(物件偵測) **1.one-stage learning** 定義: 物件偵測辨識一步到位  ==ex:Google在2015年12月提出Single Shot Detector (SSD),Google在文章摘要第一句話就寫「We present a method for detecting objects in images using a single deep neural network.」,一個深度神經網路就可以做完所有的物件偵測。== 缺點: 辨識精度(precision)低(**但還可接受**) 優點: 速度快 結論: 行動裝置仍大部分使用one-stage 例子: YOLO(YOU ONLY LOOK ONCE)->YOLOv1 v2 v3 ,Tiny YOLO Single Shot Detector(SSD) RetinaNet EfficientDet **2.two-stage learning** 定義: 分 **(1)選出物件** **(2)物件辨識** **First step** 選出物件(Region Proposal) **Second step** 物件辨識(Recognition)-->(1)物件分類(classification only)(2)特徵擷取(feature extraction)加分類(classification)  缺點: 耗時 若一張圖內有非常多物件需要擷取分類,一個物件假設需0.1sec擷取,除非有很強的gpu,不然還是無法即時運算,所以需要one-stage  優點: 精確度較高 例子: R-CNN->fast R-CNN->faster R-CNN  High-level architecture of R-CNN (top) and Fast R-CNN (bottom) object detection.
×
Sign in
Email
Password
Forgot password
or
By clicking below, you agree to our
terms of service
.
Sign in via Facebook
Sign in via Twitter
Sign in via GitHub
Sign in via Dropbox
Sign in with Wallet
Wallet (
)
Connect another wallet
New to HackMD?
Sign up