4.9
40,516 個評分
5,372 條評論

## 課程概述

In the fourth course of the Deep Learning Specialization, you will understand how computer vision has evolved and become familiar with its exciting applications such as autonomous driving, face recognition, reading radiology images, and more. By the end, you will be able to build a convolutional neural network, including recent variations such as residual networks; apply convolutional networks to visual detection and recognition tasks; and use neural style transfer to generate art and apply these algorithms to a variety of image, video, and other 2D or 3D data. The Deep Learning Specialization is our foundational program that will help you understand the capabilities, challenges, and consequences of deep learning and prepare you to participate in the development of leading-edge AI technology. It provides a pathway for you to gain the knowledge and skills to apply machine learning to your work, level up your technical career, and take the definitive step in the world of AI....

## 熱門審閱

OA

2020年9月3日

Great course. Easy to understand and with very synthetized information on the most relevant topics, even though some videos repeat information due to wrong edition, everything is still understandable.

RS

2019年12月11日

Great Course Overall\n\nOne thing is that some videos are not edited properly so Andrew repeats the same thing, again and again, other than that great and simple explanation of such complicated tasks.

## 26 - Convolutional Neural Networks 的 50 個評論（共 5,347 個）

2020年5月24日

Really an amazing course about CNN's. what an amazing instructor Andrew is. Totally recommended course those who want to learn CNN's from basic.

2019年4月27日

I think it's a good idea to remove repeated parts in the videos. Also, put all pieces toguether to give a better overview of the object detection solution

2018年4月30日

IoU validation problem is known but nothing as been done to resolv it

video editing problem

unreadable formula in python notebook for art generation (exemple :

$$J_{style}^{[l]}(S,G) = \frac{1}{4 \times {n_C}^2 \times (n_H \times n_W)^2} \sum _{i=1}^{n_C}\sum_{j=1}^{n_C}(G^{(S)}_{ij} - G^{(G)}_{ij})^2\tag{2}$$

What append ? that was great so far... =(

2019年7月8日

First of all, I thank Professor Andrew Ng for offering this high quality "Deep Learning" specialization. This specialization helped me overall to gain a solid fundamentals and strong intuition about building blocks of Neural Networks. I'm looking forward to have a next level course on top of this track. Thanks again, Sir!

I strongly recommend this specialization for anyone who wish get their hands dirty and wants to understand what really happens under the hood of Neural networks with some curiosity.

Some of the key factors that differentiate this specialization from other specialization course:

1. Concepts are laid from ground up (i.e you to got to build models using basic numpy/pandas/python and then all the way up using tensorflow and keras etc)

2. Programming Assignments at end of each week on every course.

3. Reference to influential research papers on each topics and guidance provided to study those articles.

4. Motivation talks from few great leaders and scientist from Deep Learning field/community.

2020年4月15日

I'm Zeyad, an undergraduate of Computer Engineering at Alexandria University in Egypt.

Taking this course really helped me to learn and study this field and also to implement it. It helped me advance in my knowledge. This course helped me defining Deep Learning field, understanding how Deep Learning could potentially impact our business and industry to write a thought leadership piece regarding use cases and industry potential of Machine Learning.

This specialization helped me identifying which aspects of Deep Learning field seem most important and relevant to us, apparently they were all important to us. Walking away with a strong foundation in where Deep Learning is going, what it does, and how to prepare for it.

Deep Learning specialization helped me achieving a good learning and knowledge about that field.

Thank you so much for offering such wonderful piece of art.

Best Regards,

2019年1月2日

A short (but cogent) overview of CNNs with a ton of references to read through and much more interesting assignments (than previous courses). I really enjoyed this course, I got a ton of exposure from it.

2019年4月22日

This is one of the best courses for CNNs. This gives a very deep understanding of the concepts and helps to understand the brains behind the CNNs and their working in application based environments.

2018年2月13日

Too much hand-holding during assignments, although still very good directions. Obviously the issue with the final programming assignment needs to be addressed. Fantastic lecture material, as always.

2019年1月1日

Excellent introductory course for CNN. The basic ideas and key components are explained clearly. Coding assingments helped me understand the algorithm to every little detail.

2022年1月28日

Awesome course. Programming assignments give a real world scenario for trying things. Overall a complete course for studying and implementing convolutional neural networks.

2022年6月11日

V​ery good and clearly understandable videos!

I don't feel that I'm learning much in the programming assignments. I'm able to solve the assignments but still feel very insecure. I could never solve any of those tasks from scratch without the line by line guidance comments of the assignments.

B​ut for a real life scenario there is no guidance...

2022年6月15日

really good instructions - i also like that the original papers etc. were referenced for additional reading

personally, i would wish for the programming exercises to be with less 'pre-defined' code - especially in W4 in the neural transfer programming exercise, there was a lot of code written already in terms of preprocessing etc

2022年1月4日

T​he programming assignments are great. However, there are too many constraints placed on the students. Many parts of the code are already provided, but in my opinion it would be more beneficial to allow the student to also complete many of these auxiliary codes.

2019年1月4日

Good content, videos have the occasional editing hiccups that also affect other courses in this specialisation. Assignments could be a little bit harder but do a reasonable job at familiarising with useful deep learning frameworks.

2019年10月9日

The course content is great, I felt link the programming assignments should have more information on running the Tensorflow sessions and (optional )information for people who are not familiar with Tensorflow would be great.

2018年6月10日

Great course - only thing keeping me from giving 5 stars is the consistent problem with the notebooks/grader.

2018年3月14日

assignment of week 3 has a bug about calculation of iou

2018年1月29日

I am a bit disappointed with this course , despite best efforts by Andrew. There is serious lack of rigor and while it is exciting to see things work , there is very little science to give us a methodical reason of why it works . In ConvNet we see the input data, a multi dimensional matrix get reduced in size using filtering and convolution operation techniques. From a mathematical point of view, this is clear and can be formalized but it is not clear why this process causes the ability to identify edges in a picture and evolve as we go deeper into the convNN to the real picture etc...

It seems to me this more like an alchemy rather then a rigorous scientific approach and this is why it was difficult to follow the exercises from the material of the course . I have to put concerted efforts to understand the literature which itself was not easy as it lacked rigorous mathematical and scientific approach ( why we have to increase the channels by multiples as we go deep into the conVNN ? etc...) . It seems to me the whole field is at its infancy with trials and errors - and more formalized approach is needed.

2022年3月30日

The programming assigments do not teach me anything. They are as simple as uncommenting some code. Maybe creating optional/extra credit assigments to really test my understanding would have helped.

Not knowing things is not a big problem but not knowing that I do not know is. And these assignments help me do exactly the latter. The lectures have good academic content. However, they can be edited to remove mistakes/repititions with little effort.

2019年8月31日

Great content, but this module gets far too buggy. The videos stutter and repeat as if they were going to be edited butt never were, and the programming exercises are so sloppy. The first exercise says, welcome to the second exercise, and congratulates you for finishing the course, even though the second assignment remains, that also says welcome to the second exercise! Loading a model hangs forever on one, and running the GAN crashes the kernel on the other. People in the forum have been complaining since at LEAST last year, and it's still buggy. This course content is great, but very shoddily put together compared to the rest. I am literally scared what week 5 will be like. Just clean it up guys. Hire an temp!

2017年12月7日

Wouldn't recommend because of the very low quality of the assignments, but I don't regret taking them because the content is great. Seriously the quality of deeplearning.ai courses is the lowest I have ever seen! Glitches in videos, wrong assignments (both notebooks and MCQs), and no valuable discussions on the forums. Too bad Prof Ng couldn't get a competent team to curate his content for him.

2019年8月19日

I know I am giving 2 stars :( but unfortunately this course was bit difficult and I don't know why Professor didn't first gave few fundamental concepts of computer vision. It's just my opinion maybe I'm wrong, maybe I'm right. But honestly we should have gone through some basic C.V. so that few students like myself can get a better understanding rather than directly diving into use of DL in CV.

2020年5月10日

boring and uninformative; could use improvement and some rehearsal before giving a lecture; boring and unorganized delivery; slides are horribly unorganized and boring; often times very confusing and hard to follow; should minimize the number of times the instructor references basic math and should use that time to motivate the concepts and applications

2021年8月8日

This course is an excellent introduction to Convolutional Neural Networks (aka CNNS, aka ConvNets). The instructor makes the material understandable while not straying away from going into the mathematics behind CNNs. This course also starts to get into some of the really cool applications of AI/ML/DL (such as facial recognition and neural style transfer). I also enjoy how this course (and the rest of the courses in the specialization) keeps a great balance between theory and application. It covers enough of the application (techniques and programming) that you could reasonably start working on a computer vision project straight out of the course, however, it still covers much of the theoretical and in-depth knowledge that you may need to know. The main problem I have with other CS MOOCs is that I sometimes feel that they either only focus on the theory or they only focus on the application (programming and engineering). In the former, you understand many of the in-depth concepts but still need to do a decent amount of learning on your own before you can start making stuff, and with the ladder, while you may have the programming knowledge to know how to program an application, you don't really understand the concepts so you have trouble solving a wide range of problems. For the most part, this course and this specialization straddle the line pretty well. It mainly focuses on the concepts but gives you enough practice that you could start on an application. Thank you so much to Andrew Ng and the team at Deeplearning.ai! This course was great! :)

2018年3月12日

This may be the most enjoyable course in the whole series so far. It is intuitive and fun, and the results are tangible. Very practical.

Inevitably, due to the complexity of CNN, we have to rely on frameworks such as TensorFlow/Keras, etc. to do the coding, and they are covered in this course as well. Not very deep, but sufficient. Wish they may pick PyTorch in the future as well.

The notebook and grading systems sometime have issues though. You may think you submitted the right data but actually the server side won't think so. Hard lessons learnt are: a) save the original ipynb before coding, so you can always rollback in case notebook messed up; b) save a checkpoint before submit, this will force saving and ensure you submitted the latest data, otherwise, it may submit incomplete data - some cells may still have very old data even you modified a lot; c) open anther local Jupyter notebook to experiment and mess around, with interactive TensorFlow exception, but pay attention to the expression with random sequences, when you call eval() the second time, they may have totally different value even you reset the seed upon each cell, eval() will invoke your expression again which will consume more data in the random sequence; d) never use iPad to complete your noetbook coding, :-).