200 - Image classification using gray-level co-occurrence matrix (GLCM) features and LGBM classifier

Code generated in the video can be downloaded from here:

Reference:

[sarveshbhandary]

I was wasting all my time looking this up elsewhere and all I had to do was watch this one video!
Thank you very much, this was really helpful!

[Sora-borl]

Trying to classified cloud types, using this is very effective thank you!

[heenagirdher]

Thank you so much Sir for these videos. Please create videos on other feature extraction methods also like LBP, LPQ etc.

[venkatesanr]

Hi Sreeni sir, Awesome explanation as usual, and waiting for the next video.

[ryoungseobkwon]

hello, I am a master course student in Korea. Thank you for nice course, I learned so much :)

[XX-vujo]

Still waiting for your dataset building for Unet models hope you can provide us with one. It can help us a lot more in understanding most of your recent videos. Thank you my friend. Keep it up.

[akashrprame]

Great content! I was in the dark for a while about how to use methods other than Deep Learning for image classification and this is really a great introduction to some of the concepts that are rarely found in recent literature.

How effective do you think these methods will be for a defect classification problem where both the classes have a considerable overlap?

[maheshpatel]

Thank you very much for unique 200 videos.. thanks a lot..

[efremyohannes]

Your videos are awesome and easily understandable keep on brother!!!

[himomhey]

Great content! Very understandable especially for me as a beginner who is interested in image classification.
In 17:49, what does the "0" in the index column mean?

[gustavojuantorena]

Congratulation for the video number 200 :)

[sophiez]

I have a question, you mentioned that size of 35*35, whether 35 means pixels? 35 pixels * 35 pixels! your video is really great! I love it very much!thanks!

[vwprgqg]

Dang, this is super helpful . Thankyou SIr

[sergiomorellmonzo]

In the video you say that it is not recommended to use GLCM for semantic sermentation. But this method is widely used in remote sensing even though it is computationally demanding. However, GLCM statistics are usually calculated from a moving window (usually kernel, 3x3, 5x5, 7x7). I spent some time trying to write a function to calculate GLCM-based texture statistics from a moving window. The goal is to create a texture image (eg dissimilarity) with the same shape (x, y) as the original input image. The dissimilarity value for each image pixel is calculated from the GLCM in a neighborhood (3x3, 5x5, ...). I relied on your code but I'm having some trouble getting my code to work. Would it be possible to contact you to solve it? I think it could be interesting material for the channel.

[umairsabir]

Hey Sreeni, thx for this wonderful topic. Is there any possibility of finding the anomalous part in the image and segment it using GLCM?
Also can you please make a video on ANOGAN for unsupervised anomaly detection to find the anomalous part within an image?

[stacymaimoon]

Hi Sreeni, this is really helpful. I wonder, do you know a method to recognize (background) landscapes on images just like face recognition or something similar? Would be extremely interesting to me! TIA.

[milindlande]

plz help me which is current classifer are used in CBIR model for better performance...

[crazycomer]

too many indices for array: array is 2-dimensional, but 3 were indexed... how to rectify this error sir?

[vishweshsingh]

sir i'm undergrad student just need a help from you . Actually i'm workin on malware analysis and after the progress i have to extract the features of grey scale images which is having malware hidden in it as well as clean . using cnn or any other methodogy can it be done ? suggest some ways so that i could extract the feature and classify

[bryanandujar]

why do we need 5 GLCM extraction? Can I just use 1 with distance of 1 and angle of 0 ?