DL-Simplified

DL-Simplified copied to clipboard

Deep Learning Simplified Repository (Proposing new issue)

:red_circle: Project Title : Selfie Recognition :red_circle: Aim : Create a DL model which will identify which one is selfie and which one is not. :red_circle: Dataset : https://www.kaggle.com/datasets/tapakah68/selfies-id-images-dataset :red_circle: Approach : Try to use 3-4 algorithms to implement the models and compare all the algorithms to find out the best fitted algorithm for the model by checking the accuracy scores. Also do not forget to do a exploratory data analysis before creating any model.

---

📍 Follow the Guidelines to Contribute in the Project :

• You need to create a separate folder named as the Project Title.
• Inside that folder, there will be four main components.
  • Images - To store the required images.
  • Dataset - To store the dataset or, information/source about the dataset.
  • Model - To store the machine learning model you've created using the dataset.
  • requirements.txt - This file will contain the required packages/libraries to run the project in other machines.
• Inside the Model folder, the README.md file must be filled up properly, with proper visualizations and conclusions.

---

:red_circle::yellow_circle: Points to Note :

• The issues will be assigned on a first come first serve basis, 1 Issue == 1 PR.
• "Issue Title" and "PR Title should be the same. Include issue number along with it.
• Follow Contributing Guidelines & Code of Conduct before start Contributing.

---

:white_check_mark: To be Mentioned while taking the issue :

• Full name :
• GitHub Profile Link :
• Email ID :
• Participant ID (if applicable):
• Approach for this Project :
• What is your participant role? (Mention the Open Source program)

---

Happy Contributing 🚀

All the best. Enjoy your open source journey ahead. 😎

I would like to work on this, frankly I don't know much but I will learn and do. I am a newbie(first year UG student). I want to learn and grow

Please do not to spam in every single issue. This might cause escalation as you are breaching Code of Conduct and Contribution Guidelines.

@rafiya2003

Full name : Siddhant Tiwari GitHub Profile Link : github.com/stiwari-ds Email ID : [email protected] Participant ID (if applicable): stiwari-ds (Quine)

Approach for this Project :

1. Implement preprocessing methods to isolate faces and segment background using OpenCV
2. Train custom CNN models with processed vs. unprocessed images to compare results
3. Use pre-trained CNN architectures for feature-extraction as well as fine-tuning
4. Experiment with Vision Transformer models for treating images as a whole without any preprocessing

What is your participant role? (Mention the Open Source program): Individual Contributor, SSOC-2023

Issue assigned to you @stiwari-ds

Full name : Atharv Pal GitHub Profile Link : https://github.com/atharv1707 Email ID : [email protected] Participant ID (if applicable): <I am new to Open Source, do tell me how to find it , I assume it's atharv1707> Approach for this Project :

Here is my suggested approach for this issue :

1. Exploratory Data Analysis (EDA): Understand the dataset's characteristics, including image distribution, resolution, and color composition.
2. Preprocessing: Standardize images for training by resizing, converting to appropriate formats, and normalizing pixel values.
3. Model Architecture Selection: Experiment with CNN architectures like VGG, ResNet, and Inception, exploring transfer learning for efficiency.
4. Model Training: Train models on a subset of the dataset, by leveraging data augmentation to prevent overfitting.
5. Evaluation: Assess model performance using accuracy, precision, recall, and F1-score metrics to identify the best algorithm.
6. Hyperparameter Tuning: Fine-tune model parameters such as learning rates and batch sizes for optimization.

What is your participant role? Contributor, GSSoC'-2024.

Well your approach looks good to me. You can start working on this issue @atharv1707. No need to worry about the participant ID, that's not applicable to you.

Hey! I tried for the past 3 hours working on this, and I have thoroughly gone through out the dataset concluded that the dataset is not normalized, with selfie and id image ratio way too skewed. Also, I believe there is somewhere in between the images are mislabelled, which may in turn require manual labelling.

I request @abhisheks008 please unassign me from this issue. There are a few other projects I would like to contribute to.

Thanks man!

Cool @atharv1707

Full Name: Aarya Prajapat Github Profile Link: https://github.com/Aarya31 Email Id - [email protected] Participant ID - Aarya31

Approach for this Project: Firstly, i would do data augmentation, standardize the images for training. As it is image classification/selection problem, i will use ResNet, inception, DenseNet for efficiency and effectiveness. Train the model and assess model performance using accuracy, precision,etc.

Hi @Aarya31 issue assigned to you.