Leukaemia Classification using DL

[achrekarom12]

Pull Request for DL-Simplified 💡

Issue Title : Leukaemia Classification using DL

• Info about the related issue (Aim of the project) : To classify normal from abnormal cell images of Leukaemia.
• Name: Om Achrekar
• GitHub ID: achrekarom12
• Email ID: [email protected]
• Idenitfy yourself: (Mention in which program you are contributing in. Eg. For a JWOC 2022 participant it's, JWOC Participant) SSoC '23

Closes: #314

Describe the add-ons or changes you've made 📃

• Data collection: From the link of the dataset given above.
• Data preprocessing: Preprocessed the image in order to have all images in equal shape.
• Model selection: Chose three Image detection architecture VGG16, ResNet50 and Inception for Image detection.
• Comparative analysis: Compared the accuracy score of all the models.

Type of change ☑️

What sort of change have you made:

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Code style update (formatting, local variables)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested? ⚙️

The evaluation metrics I used to assess the models:

• Accuracy
• Loss

Checklist: ☑️

• My code follows the guidelines of this project.
• I have performed a self-review of my own code.
• I have commented my code, particularly wherever it was hard to understand.
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• I have added things that prove my fix is effective or that my feature works.
• Any dependent changes have been merged and published in downstream modules.

[github-actions]

Our team will soon review your PR. Thanks @achrekarom12 :)

[achrekarom12]

@abhisheks008
I definitely agree that the approaches I used have quite a low accuracy between 65-70. This is because I have use only 1/3rd of the available data. The image processing were little difficult here because it was exhausting the whole RAM available. There were over a 10K images to be processed. In the notebook I have also mentioned the approach while taking whole dataset into consideration but commented it out. Please let me know if there's anything more I can do.
Thankyou!

[abhisheks008]

Anything else you can apply on this dataset? Accuracy of 70% is not less, but the standard you have fixed for the contributions, it seems below PAR accuracy.

Try to fix the accuracy atleast to 85-90 mark.
@achrekarom12

[achrekarom12]

I hope it would be good enough. Although I couldn't get near to 85 but overall the accuracy has been improved remarkably with highest performing model showing the accuracy of 80.20%. Please let me know if any further improvements you need in this. Thankyou!

[abhisheks008]

Your PR is approved and ready to be merged.
@achrekarom12