Facial Expression Recognition with PytTorch

[sreevidya-16]

Deep Learning Simplified Repository (Proposing new issue)

:red_circle: Project Title : Facial Expression Recognition with PytTorch :red_circle: Aim : To develop a facial expression recognition system using PyTorch that classifies facial expressions into various categories using a convolutional neural network (CNN).
:red_circle: Dataset : The dataset used for this project is available on Kaggle. It contains images of faces categorized by different expressions such as happy, sad, angry, surprised, etc. URL : https://www.kaggle.com/code/veb101/facial-expression-recognition-using-pytorch
:red_circle: Approach :

1. Exploratory Data Analysis (EDA): Perform EDA to understand the dataset, visualize the distribution of different expressions, and preprocess the data.
2. Model Development: Implement and compare 3-4 different algorithms/models to identify the best-performing model based on accuracy scores.
3. Model Evaluation: Evaluate the performance of each model using appropriate metrics and select the best-fit model.
4. Visualization: Provide visualizations of the results and model performance.

---

:white_check_mark: To be Mentioned while taking the issue :

• Full name : Patamsetti Sree Vidya

• GitHub Profile Link : https://github.com/sreevidya-16

• Email ID : sreevidyapatamsetti@gmail.com

• Approach for this Project :

  1. Exploratory Data Analysis (EDA):

• Load and inspect the dataset.

• Visualize the distribution of facial expressions.

• Display sample images for each expression category.

• Clean the data if necessary.

  2. Data Preprocessing:

• Resize images to a consistent size.

• Normalize pixel values.

• Apply data augmentation using Albumentations (rotations, flips, zooms).

  3. Model Development:

• Model 1: Basic CNN with convolutional and pooling layers.

• Model 2: Fine-tune a pre-trained CNN (e.g., ResNet50) from the TIMM library.

• Model 3: Use another architecture (e.g., VGG16) and fine-tune.

• Compare models based on accuracy scores.

  4. Model Evaluation:

• Evaluate each model using accuracy, precision, recall, and F1-score.

• Select the best-performing model.

  5. Visualization:

• Visualize model performance and comparison.

• Provide conclusions and insights.

• What is your participant role? (Mention the Open Source program? I am Participating in GirlScript Summer of Code 2024 (GSSoC'24).

[github-actions[bot]]

Thank you for creating this issue! We'll look into it as soon as possible. Your contributions are highly appreciated! 😊

[abhisheks008]

Already present in the repository. Closing this issue as duplicate problem statement.