Это видео недоступно.
Сожалеем об этом.
HOG Intuition | Simple Explanation | Feature Descriptor & Engineering
HTML-код
- Опубликовано: 12 июн 2021
- 🔥 Simplest explanation of Histogram of Oriented Gradients (HOG) & building HOG representation for real image data. We will understand the high-level logic through an example in this short video. Happy learning!
🔥 At Skillcate, we give you hands-on experience of industry ready projects having real-life complexity and solve them end-to-end for you.
Along the way, we will share Do-It-Yourself kits with you, having:
-- 1. dummy business case,
-- 2. problem dataset,
-- 3. python code,
-- 4. business analysis tools,
for you to solve these problems yourself.
Needless to say, you may always reach out to us for any help you may need to achieve your objectives, by commenting below or through our contact details provided in the description part below.
Do like, share & subscribe to our channel.
🔥 Keep in touch:
Website: www.skillcate.com
LinkedIn: / 67209084
Email: skillcate@gmail.com
Hey guys!! Glad to see such amazing feedback on this ML Project🤗 Need your support in reaching out to more learners by subscribing to my channel 🙂 Also, join me on my Skillcate Discord Server: discord.gg/GyMBfD4ER5 🙂 Let's talk Machine Learning ❤❤
Definitely would recommend this as the video to watch if it's your first time trying to understand HOG. Best explanation I've seen so far! Thank you.
It is the best explanation video for HOG i have come across. Simple and clear. Thanks.
Wow, thanks!
Amazing and underrated video :)
Thanks 🙏🏼
So at first, we are going to make a matrix from our three R,G,B matrices and then we do these calculations on it ?
Dear Ahmad, that's an excellent question!! As we already know, at every pixel, the gradient has a magnitude and a direction. For color images, the gradients of the three channels are evaluate. The magnitude of gradient at a pixel is the maximum of the magnitude of gradients of the three channels, and the angle is the angle corresponding to the maximum gradient.
So, HOG treats the three RGB channels/matrices separately for magnitude + orientation computation and then takes the max of those values across the three channels.
@@skillcate Thanks for clearing this all :)
Can you please share me how the Pixel takes the values in part of calculate gradient and orientation
And why you select only 85 for highlighted part
What is process going on behind that.
How matrix is formed and how it takes the values?
Hey Pranali, I selected '85' randomly. Gradient & Orientations are calculated for all values in the shown matrix.
In simple terms, idea is to check for pixel intensity changes from left-to-right (x-direction) and bottom-to-top (y-direction) of a selected pixel. Gradient values tend to shoot up near boundary of objects, as colour intensities change drastically along boundaries. If you go to 3:15, you will see how HOG representation captures only the object boundary information, smartly discarding the rest.
And once we have our gradient values (Gx and Gy), calculating magnitude and orientation are simple Pythagoras Theorem.
Hope this answers your question :-)
This is a great video!! Help me a lot. Thank you!!
Glad it helped!
helps alot
can you give a link of an article which provides a simplified information for a beginner
www.analyticsvidhya.com/blog/2019/09/feature-engineering-images-introduction-hog-feature-descriptor/
Can i have these slides???