Azure’s Custom Vision — ML Prediction

→ Guys this is the continuation of the previous video team ML Training.

→ Basically what we have done in the previous article is ML Training.

→ ML Training in short is if we want to extract data from a random set of images having some similar thing — like a fruit in it or a bank’s deposit form.

→ What we do next is upload multiple sets of images in variation and tag them using Azure Custom Vision Trainings.

→ Now the fun part is prediction. Now your ML knows about the data. Basically, one model in ML is for one specific work

eg. identifying fruits or deposit form

ML Training:

→ In training what we do is to upload multiple images having some similar properties.

→ Once Uploaded, we tag each image. The more the tag, the better ML prediction (i.e. extraction ).

→ Each model represent = Specific work =

→ Check here for the ML Trainings article and video

ML Predictions

What is Prediction?

→ Once have our model ready we can upload as many documents to identify or extract data.

→ The effort now of manually copying the data from the physical paper or softcopy file is not required and time and effort are saved.

→ Let us go to the models we have created

→ Let us now see the advantage of Machine Learning:

→ In the previous video, we uploaded multiple pomegranate videos and trained them.

→ Now our machine knows pomegranate and can understand if that fruit is there in any image

→ Let us open Azure Cloud Portal.

→ Search for Custom Vision Service

→ We now see the models we created. Click on the first one.

→ Let us now open Custom Vision Portal

→ A new tab opens. Here we will our models. Click on the model trained. Say the second one.

→ Click on the Quick Test button

→ Now File Explorer will come. Here we can upload any image that contains Pomegranate.

→ Let us see if ML can detect the fruit.

→ Let us try with another image. Click Browse local file.

→ Now any photograph with pomegranate fruit can easily be detected by ML.

→ Now think how powerful can this be. If we get documents in bulk or anything that we can use ML can easily identify and separate images.

→ Being a Javascript developer and using ML in my backend with Node.js is fun and powerful for applications.

→ Will highly recommend exploring and feel free to ask queries.

Closing Thoughts:

We have learned how to use Machine Learning to identify the objects in random images using ML Prediction.

Being a developer myself and using it in my project, applications become more intelligent with these features when dealing with bulk documents or images.

Thank you for reading till the end 🙌 . If you enjoyed this article or learned something new, support me by clicking the share button below to reach more people and/or give me a follow on Twitter and subscribe Happy Learnings !! to see some other tips, articles, and things I learn about and share there.