Learn more about Python deep learning recently, because you want to use Python to do graphics recognition and get the relevant introductory books. Chinese is the best.
is to give a picture that identifies what the image is.
Reply content:
This is a
a more completeLearning path for image recognition using deep learning, not deep learning
Shortcut！
1. Pattern Recognition
Judging from the language of your problem, the master seems to have no higher understanding of pattern recognition.
Therefore, before doing the image recognition based on deep learning, it is recommended to read the pattern recognition and computer vision related books first. First
Understanding the image of the information itselfTo try to identify.
"Learning OpenCV", I think it is a very good introductory book on the above fields, and there is also a Python interface.
2. Machine learning
It is not clear how much you know about machine learning and its related fields. Before learning the theory of deep learning, it is suggested to learn the shallow model and its theory. Of course there are no particularly good Chinese books. But "machine learning", "statistical learning method" is still worth a look. Foreign language Recommendation "Pattern Recognition and machine learning" and
"Machine learning:a Probabilistic Perspective", the latter containing the chapters of the Deep Neural network。
3. Deep learning
Finally deep learning in Python is the Theano of the University of Montreal (Welcome-theano 0.6 Documentation
）。 It has a corresponding English version of the Deep Learning course (contents-deeplearning 0.1 documentation
), will involve
extension of shallow model to depth model。
I have translated some of the documents into Chinese and put them on GitHub. Poor quality, if interested can join together (syndrome777/deeplearningtutorial GitHub
）。
There is also a Stanford document, mainly on automatic encoder, Chinese scholars have completed the translation (UFLDL tutorial-UFLDL
）。
Deep learning is very hot, making some students eager to join. But I still suggest that we first have the basis of pattern recognition and machine learning to learn later, so
it will be compared to the force after the late play。 Deep learning is a relatively fire topic of machine learning, and machine learning is a direction of computer science, a cross-discipline of computer science and statistics. And Python is a computer programming language.
So in theory python can implement any algorithm, including deep learning algorithms. The deep learning algorithm can also be implemented by any computer language. So the question asked by the Lord itself is problematic.
I guess the main idea is to understand what Python's algorithm packages are. Python Machine learning algorithm package recommended Scikit-learn:machine learning in Python
, CV-related libraries are more well-known and the document is more complete is OPENCV.
As for deep learning, the main problem is to understand the most basic machine learning algorithm for good, and then can look at various tutorial and paper, in essence, neural network. Python's best in-depth learning material, not one.
Deep Learning Tutorials
Programming computer Vision with Python:techniques and Libraries for Imaging and retrieving information
@issac Syndrome's answer was more complete. Here are two additional information on deep learning:
- Hinton in Coursera's neural network course:https://www. Coursera.org/course/neu ralnets
- On the other hand, if you do deep learning, you may need to use GPU parallel computing, now the most popular GPU computing framework is cuda ( https://developer.nvidia.com/cu da-downloads ), which is also the support technology behind Theano.
Stanford recently opened a course to introduce convolutional neural networks (a commonly used deep learning model), and the job is also based on Python, and the main interest of the topic is to be followed.
Stanford University cs231n:convolutional Neural Networks for Visual recognition
