Skip to content
Permalink
Browse files

Committer: 陈云 <chenyun@s-MacBook-puro.local>

 Changes to be committed:
	modified:   book/learning_materials/80_advanced_perception/10_object_detection.md
  • Loading branch information...
陈云 陈云
陈云 authored and 陈云 committed Dec 20, 2018
1 parent 0c60460 commit 5df3c1973d758e03aefe1acf172ee3c95a96805e
@@ -1,4 +1,11 @@
# Object detection {#object_detection status=draft}

Assigned: Nick and David

Assigned: Yun

## Background

Object detection helps autonomous vehicles detect different objects. The difference between object detection and classification is that detection algorithms not only output the class labels that the objects belongs to, but also output the bounding boxes for the objects.

Generally, object detection can be
@@ -10,7 +10,7 @@ Traditional ways of object classification extract features or descriptors for di

An object classification algorithm takes images as input and output what objects it contains. For simplification, HOG plus two-class (binary) classifiers are explained because it has the best performance in traditional algorithms.

## Steps in object classification
## Traditional Steps in object classification

The main pipeline is shown as following:
<figure class="stretch">
@@ -34,7 +34,9 @@ The general principle is that learning algorithms treat feature vectors as point

Support Vector Machine (SVM) is one of the most popular supervised binary classification algorithm. It tries to find the hyperplane that can maximize the margin of two classes. It is also widly used combined with HOG for better accuracy in object classification.

## Deep Learning in object classification

Except traditional methods, deep learning approaches perform an astoundingly better accuracy in computer vision. Details will be illustrated more in the next module of object detection because it includes the task of object classification.



0 comments on commit 5df3c19

Please sign in to comment.
You can’t perform that action at this time.