Tutorial 26- Create Image Dataset using Data Augmentation using Keras-Deep Learning-Data Science

Please join as a member in my channel to get additional benefits like materials in Data Science, live streaming for Members and many more
/ @krishnaik06 In this video we will understand about the max pooling layer in CNN
Please do subscribe my other channel too
/ @krishnaikhindi
Keras blog: blog.keras.io/...
Connect with me here:
Twitter: / krishnaik06
facebook: / krishnaik06
Instagram: / krishnaik06

[@Prachi_Mayank]

thank u thank u for always helping me like an angel... u are just great sir..keep it up... people like us who are new in this field need your support, so thank u again sir.

[@nurafiqahmahizan]

simple and comprehensive explanation. thank you sir!

[@mertceylan2080]

Short and brilliant tutorial ! Thank you so much Krish... May Allah bless you

[@arjunpukale3310]

Sir please upload a video on how to convert image dataset stored in directory as numpy array with image augmentation

[@SAURAVkumar-sg8mk]

thankyou sir,,,it helped me in data generation for my CNN model

[@alluprasad5976]

Thank you Krish..do more on segmentation and detection

[@cool_videos6016]

How can you data augmentation on selected image classes when a dataset is imbalanced

[@pawansran4690]

This video has been really helpful. Thank you so much

[@vishaljhaveri7565]

Thank you, Krish Sir. It was a good tutorial.

[@talasrour5068]

please i need to know how can apply the same technique to a dataset of 258 image and generate 20 copies from each single image and save it and achieving 1000 as total

[@sandipansarkar9211]

That was a superb video.Just want to get my hands dirty with practise.Thanks Krish.

[@ausheekhamesra6583]

Just wondering how I could change the names of the augmented images that are being saved? For example instead of the randomized numbers "cat_0_1449.jpeg", how could I save it as cat_1, cat_2, cat_3.... Thank you

[@movsesghazaryan1020]

is there a way of using those images without saving them on hard drive? edit: you can use .fit() . You need to give datagen.flow(...) as an argument. For anyone looking for the same question's answer.

[@heenagirdher6443]

Thank you so much sir for such a great tutorial

[@zebafariha8312]

Sir, can you please show how can I augment data on multiple images(folder of certain labelled pic) to make imbalanced data a balanced one. I am struggling with that

[@saibaba2962]

Dear Sir, I am doing research on healthcare and have an MRI dataset. I need to perform data augmentation on MRI images. Could you please suggest any tools or Python code for MRI data augmentation? Thank you.

[@sahubiswajit1996]

Why we are doing this line? x = x.reshape((1,) + x.shape) # this is a Numpy array with shape (1, 3, 150, 150)

[@abhishekbordoloi9923]

Even I am stuck in this. Did you get it?

[@avneshdarsh9880]

1,3 is colour code ,150*150 size of image

[@shindepratibha31]

Because ImageDataGenerator takes data in that form where, 1----number of images 3----number of channels 150x150---image size

[@teacherHub6723]

@@avneshdarsh9880 1---batch size 3---channel ( RGB ) and 150*150

[@aryansingh2105]

Sir if possible please make a video on how to create a synthetic categorical dataset for generating a CSV file. I have watched your video in which you have used DataFrame for the generation of artificial data, but attributes in that dataset have no correlation. Thank you.🙏🏻

[@kin_1997]

love your content man very helpfull

[@Motivational-quotes143]

How can we increase the dataset for numeric dataset?

[@relaxpeacefulmind2440]

In this augmentation if we do all augmentations at a time won't it mix everything , means it will flip as well as rotate as well as elongate in single image

[@shrinivas8035]

so here we have to store different images in separate folders and similarly store results in separate folders and then use the new data set for the learning algorithms right?

[@xXMaDGaMeR]

do you have to save it do local drive ? or could you possibly pass the augmented data within the training dataset ?

[@pruthvirudraswamimath860]

Hello sir I want anemia prediction from eye image project and how it works plz sir

[@jeamsdere9636]

thank sir, it was great tutorial

[@wolfinstincts29]

the fourth dimension is for batch number

[@mahenderboda9911]

Let's say My problem is not classification but image to image problem where both the input and output are both images. and I want to expand my data. In this case how can I apply imagedatagenerator to get the same transformations applied to both the input images and output images.????? Thanks in advance

[@VictorVelazquezEspitia]

Excelent video, thank you

[@imanefahim8557]

Nice video, I did data augmentation "on fly" inside my model, any ide how to visualize it?

[@satyaganesh7159]

sir please create a tutorial on MNIST dataset with indepth intution, which will be useful

[@sourinroy992]

is data augmentation only applicable to image datasets? what about other types of datasets?

[@mujeebrahman5282]

Sourin Roy yes as far as I am concerned. I may be wrong.

[@rumitpatel884]

in other type you can use like a undersampling and oversampling concepts

[@wajeehanaz9115]

Sir i want to increase image dataset using smote technique ...please tell me how to do it???

[@NomanNawaz-i9j]

how to change the images save location after augmentation , i want to save images after augmentation in my pc drive ?

[@siddhantkumar698]

Can we use GAN to generate images for CNN

[@tirthadatta7368]

Sir, can u please make video on data augmentation using multiple image and how to split image data into train and test??

[@vandanapal7321]

i haven't used kaggle dataset and my images are no labelled like an image is a cat or dog ....what condition should i apply in the if else block for appending 1 if a dog and 0 if cat

[@SoumyaPanigrahi-wt7il]

how to create preview folder and use if i am using colab?????????????????????/

[@satinder_sartaaj]

How can we balance image dataset by using this augmentation??

[@codingtamilanshorts]

Data argumantation for MATLAB code of CNN is available ??

[@everythingisinfrared6029]

Sir, is image augmentation applied only at the input image (e.g. the image that shows the cat or dog), or also on the label image (binary image)? Thank you.

[@rutveek6050]

getting bluish images ? how to solve?

[@jayasandeepreddy3474]

Hi Krish, after augmenting the images when we pass the data for training , do we need to maintain height and width of all images same ?

[@louerleseigneur4532]

Thanks Krish

[@mianshani9115]

How thia data set used in website??? Such as i want to make data set for Hat pics and put that pic on real time face detection..? Could you tell me how to use?

[@foruniversity2525]

thank you very much

[@wanianawaz6999]

Could not import PIL.Image. The use of `load_img` requires PIL. sir this error occur continiously , how we can install pil in jupyter?

[@christopherami8142]

same here ?

[@kamleshsahane9439]

pip install pillow

[@saniaurooj5450]

your video is very helpful. Can you tell me which data augmentation technique is used for time frequency domain spectrogram (images) ?

[@unknownfacts3716]

don't go deeper into this topic, just cover the basic one

[@notgamer8802]

which link did you implement program right now?

[@yughendermeda6231]

Thank You Sir

[@ZulfiqarAli-cg5qw]

I need help ..Mara error aarha haii save to dir py... augmentation naii show ho rahii preview ma mif someone knw plz help ma

[@abhinavgarg5611]

For those facing this error FileNotFoundError: [Errno 2] No such file or directory create a folder named preview in the same path where your notebook or py file is located and then you will be able to see augmented images in this folder

[@Jaydeep-ve4jy]

Hi Krish, thank you for the amazing video. Is there a way we can also preserve labelled data while performing image data augmentation?

[@mmskawish]

If you have corresponding .xml files, then "Roboflow" (google it plz) may work for you. Let me know if you need further help.

[@nishatchowdhury1466]

Is it applicable for machine learning??

[@rajabalimondal7308]

Sir please upload a video on image pre processing for model fit()

[@aisyahrobi3623]

FileNotFoundError: [Errno 2] No such file or directory...can someone help me with this... I really need help for my project

[@shubhajitroy1787]

Use full path of the preview folder in place of save_to_dir='preview' it will work.

[@andrmtts]

thank you!

[@vjukulkarni6057]

How if we have n numbers of images

[@forever-fz1hk]

krish sir just one question please answer....to speed up my processing iam practising them on kaggle as it provides much fast speed....but please answer how to train a image from my computer...or how to set path

[@krishnaik06]

U probably cannot do that

[@subhadeepdutta5501]

If I have .nii file and how to augment based on CT, MR

[@9-se-5]

Hi sir getting this error please help em out Input data in `NumpyArrayIterator` should have rank 4. You passed an array with shape', (2720, 4106, 3))

[@iliasp4275]

you must have forgoten to put this line x = x.reshape((1,) + x.shape) at 3:41 to put and 1 in front of it to make it rank 4

[@ajanthakaipu1790]

Sir please explain graph convolution neural network

[@arshad1781]

nice

[@tausifahmed751]

MANKIND'S ANGEL YOU ARE, YOU ARE MY SAVIOUR. INDIA ROXX

[@wanianawaz6999]

i am stuck with so many error please help me out

[@hinanaz4254]

i have getting error any one can help us

[@nidhikundu3620]

how to do data augmentation of multiple images together??

[@gujjalapatiraju7435]

Loop through the images you have in a folder and apply this on each image.

[@divyarajvs8890]

FileNotFoundError: [Errno 2] No such file or directory: is the error i am getting even i though i copy the correct path from my system where the image file exists

[@Omarismail-vs4jl]

make sure you place it in the same directory your notebook is in. for example Anaconda3/cat.jpg

[@aisyahrobi3623]

Hello, I also have the same error as you....how you solve it?

[@mianshani9115]

Source code link??

[@EngRiadAlmadani]

Good jop

[@Kumar08]

Please provide Github link for this code.

[@prasunprakash2297]

Please provide the github link for this code

[@priyankadinda6670]

where to create the 'preview folder.

[@divyarajvs8890]

FileNotFoundError: [Errno 2] No such file or directory:

[@sksujanislam3906]

@@divyarajvs8890 create the preview folder in the same location where the notebook file is located.

[@renarayana4177]

Can you share the code???

[@nishatchowdhury1466]

I have 90 real datasets..bt i need total 1000 dataset..shall i apply this technique for machine learning??

[@prajwalbanagar3179]

I have 30 and have CNN to do😆

[@darpan810]

same technique I applied but showing error "Could not import PIL.Image. The use of `load_img` requires PIL." Please help sir

[@santhanalakshmism22]

Pip install pillow

[@darpan810]

@@santhanalakshmism22 will try thanks

[@santhanalakshmism22]

@@darpan810 My pleasure

[@darpan810]

@@santhanalakshmism22 getting same error after this run command

[@santhanalakshmism22]

Did u install pillow ? Please check the version

[@ranjithkumarsamala6416]

dude, please do post code in descritption

[@aravindnaidu4819]

blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html

[@saurabhkumar289]

Please send this code

[@aravindnaidu4819]

blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html

[@shreenivaschityala3699]

Indian Andrew Ng

[@khankhn4285]

I want to increase (augment) the number of images that i have (50 images) using ImageDataGeneator. I found the following code, but it only augments one image per time (one file). I am new to python, so, if there any other easier method to automatically augment the 50 images at once and save the outcome in a new folder... Any solution

[@khankhn4285]

import numpy as np import keras from keras.preprocessing.image import ImageDataGenerator, array_to_img, img_to_array, load_img datagen = ImageDataGenerator(rotation_range =15, width_shift_range = 0.2, height_shift_range = 0.2, rescale=1./255, shear_range=0.2, zoom_range=0.2, horizontal_flip = True, fill_mode = 'nearest', data_format='channels_last', brightness_range=[0.5, 1.5]) # This is my problem, It loads only one file, and i am searching for an automated method to load number of files together and save them in another folder. img = load_img(r"C:\Users\user 1\Pictures\people_1\1.jpg") x = img_to_array(img) x = x.reshape((1,) + x.shape) i = 0 for batch in datagen.flow (x, batch_size=1, save_to_dir =r'C:\Users\user 1\Pictures\people_1\preview', save_prefix ='people2', save_format='jpg'): i+=1 if i>10: break

[@khankhn4285]

The expected result from the above code is 10 augmented images extracted from the main file of 1.jpg, while I have another 50 files, I am searching for a quicker method to augment them all in one code

[@sonamkori8169]

Thank You Sir