How To Collect Images For Training Neural Network in just 5 minutes?
This is foremost question to every Deep Learning aspirants…How to collect hundreds and thousands of images for training your deep learning model.
Let’s say, you want to create a model that classifies dog and a cat.One may think that.. I will download it manually..! 😱Like Really !! If we go with this way we would end up being collected maximum 100 -200 images and those images may noise ,irrelevant sometimes.Also,doing this task manually will take much much time.
One important thing is if we fed these noisy 100–200 images to our model,our model will overfit (means it will learn the dataset)which results in low accuracy .
So..Not to worry ! I am here to share with you how you can reduce your headache and do this tasks in just 5 minutes.
Get ready!! Here are steps that you need to follow?
STEP 1: Go to Google extensions and search for “Fatkun Batch Image Downloader extension and add it to your extensions once you add that extension it will see something like this.
STEP 2 :Go to https://www.flickr.com/ . I find this amazing webite where you can search for anything you want and you will get absolutely less noisy and relevent images.
STEP 3 :Load images as many you want and press at this and click on Download [Current Tab].
this will load all that images which are present at the flickr.com page.
Step 4 :Before this..Make sure you have enable automatic download in google settings (by doing this it wont ask you every time).Now..At very first time,Name the image which ever you want to here i am taking dog images.
And click on download .This will download those images serially in your downloaded folder
Wohhh!!!! Thats So Great ! Isn't it ? .
Congratulations!!!!!!!! You've succesfully created a dataset for your deep learning model.
Thank You For Reading!!
Have a Good Day ! Be happy and Keep Learning !
