search

goodboychan / goodboychan.github.io

jupyter blog
Apache License 2.0
129 stars 115 forks source link

Advanced Operations, Detecting Faces and Features | Chan`s Jupyter #31

Closed utterances-bot closed 4 years ago

utterances-bot commented 4 years ago

Advanced Operations, Detecting Faces and Features | Chan`s Jupyter

After completing this chapter, you will have a deeper knowledge of image processing as you will be able to detect edges, corners, and even faces! You will learn how to detect not just front faces but also face profiles, cat, or dogs. You will apply your skills to more complex real-world applications. Learn to master several widely used image processing techniques with very few lines of code! This is the Summary of lecture “Image Processing in Python”, via datacamp.

https://goodboychan.github.io/chans_jupyter/python/datacamp/vision/2020/08/02/03-Advanced-Operations-Detecting-Faces-and-Features.html

geosanti088 commented 4 years ago

In Privacy protection, you arent applying gaussian filter when mergin, your photo "Blurred faces" photo is a "face" cropped photo. I think the code has some problem, because I get the same result and seems that merge isnt working

goodboychan commented 4 years ago

Hi,

Thanks for the comment. I didn't notice that is wrong. I updated the example in that section. The reason why black image is shown that the output for gaussian filter is normalized. So the range of output is from 0 to 1, and after merging it, it will be shown as black. To avoid this, one argument is required as follows:

blurred_face = gaussian(face, multichannel=True, sigma=8, preserve_range=True)
geosanti088 commented 4 years ago

Thank you, very much. I tryed to fix it yesterday for many hours and now you make my day. 👍