Data Collection for Image Classification using Python - Google Images

[@kurrysamir]

This was God sent Srivatsan Sir. Was struggling to get a dataset I was building for a Custom classifier to identify furniture, in particular Tables, Beds and Chairs. This is as a part of IBM Applied AI Coursera course that I am doing. I was able to get the datasets for furnitures and was able to classify them with confidence scores between 0.88 and 0.9. Super thanks :)

[@AIEngineeringLife]

Good to know and glad it was helpful. If u need more images I have a video coming tomorrow on it

[@kurrysamir]

@@AIEngineeringLife awaiting for it. Will surely refer.

[@TejasPhase]

Thanks, @AIEngineering Sir, for this video. It really helps in building "Custom Image Classification". I have searched like this for a long time but not found any appropriate way. Will experiment with it and let you know if raised any problem. Thanks once again.

[@AIEngineeringLife]

👍

[@zaldanadaniel]

Amazing!. Tank you!

[@onurdemir7173]

I get an error when I run the code '!ls download' in step 4 Error: 'ls' is not recognized as an internal or external command, operable program or batch file. If you have any idea then it will be helpful. Thanks in advance..!

[@dhruvipatel4853]

I am getting this error while downloading, I have tried a few things but it's not working for me. If you have any idea then it will be helpful. Thanks in advance..! Error: Evaluating... Could not open URL. Please check your internet connection and/or ssl settings If you are using proxy, make sure your proxy settings is configured correctly An exception has occurred, use %tb to see the full traceback. SystemExit C:\Users\hp\Anaconda3\lib\site-packages\IPython\core\interactiveshell.py:3334: UserWarning: To exit: use 'exit', 'quit', or Ctrl-D. warn("To exit: use 'exit', 'quit', or Ctrl-D.", stacklevel=1)

[@AIEngineeringLife]

Have you checked if you are not beyond any firewall that is blocking it?.. If so can you try my other video on microsoft bing image search and see.. Just want to rule out the access is not getting blocked

[@dhruvipatel4853]

@@AIEngineeringLife I tried after disabling the firewall but it's not working. With microsoft bing image search I am able to download the images. Thank you for your suggestions.

[@sdhilip]

Informative and nicely explained. If possible, please share the code. Thanks for sharing and posting

[@AIEngineeringLife]

Thank you.. Here it is - github.com/srivatsan88/KZbinLI/blob/master/Data_Collection_for_Image_Classification_Google_Search.ipynb

[@sdhilip]

@@AIEngineeringLife Thank you. Is it possible to download more than 100 images? I tried but not successful. Then I google it, come to know that google_images_download.py set limit into 100. Then I have increased it but not working. Any suggestions?

[@AIEngineeringLife]

@@sdhilip .. You have to do it in batches.. so download 100s in loop.. i will talk about another way this week where limit is higher but does not use google image. Google search has better search result though

[@sdhilip]

@@AIEngineeringLife Thanks a lot for your feedback

[@flamboyantperson5936]

please make more videos on pyspark machine learning

[@AIEngineeringLife]

Anything in Specific in pyspark ML?.. I already have some video covering end to end ML using pyspark, topic models and Spark H2O. I have one for time series coming. Any specific content you are looking for in pyspark?

[@flamboyantperson5936]

@@AIEngineeringLife I am looking for anomaly detecting using iforest algorithm in pyspark esp if it's banking data but you can take any dataset of your choice. I have just started watching your videos and they are awesome. I loved your content. Thank you.

[@AIEngineeringLife]

Have you seen Linkedin Spark isolation forest package.. but it is only in scala engineering.linkedin.com/blog/2019/isolation-forest Having worked on banking data a lot I have used many times xgboost for detecting frauds if we have labeled data but if not there is not much option

[@flamboyantperson5936]

@@AIEngineeringLife Thank you for sharing. I am looking in pyspark I don't know scala. If you have time could you make in pyspark?

[@AIEngineeringLife]

Let me try as in case if isolation forest does not have pyspark version then distributing it will be difficult. We have to set it up as python udf but again scaling across records it might not help. Will check it out and see