search

Recode-Hive / machine-learning-repos

A curated list of awesome machine learning frameworks, libraries and software (by language). I
MIT License
55 stars 126 forks source link

Title :- Pharmaceutical Drugs and Vitamins Images Detection #221

Closed aaradhyasinghgaur closed 10 hours ago

aaradhyasinghgaur commented 4 weeks ago

Description :- The below dataset includes images of popular pharmaceutical drugs and vitamins in the Philippines. This is dataset can be used for classifying drug images using CNN and transfer learning. Currently, there are ten available classes of pill images. One of the primary uses would be to assist individuals in identifying medications and vitamins. This could be helpful for patients who have multiple prescriptions or for caregivers who need to manage medications for others. By taking a picture of the medication, the system could provide information about its name, dosage, and usage instructions.

Datset I'll use :- https://www.kaggle.com/datasets/vencerlanz09/pharmaceutical-drugs-and-vitamins-dataset-v2

Solution I propose :-

  1. Utilizing 5 models such as DenseNet121 , Xception, VGG16, ResNet50, and InceptionV3 for image classification.
  2. Applying data augmentation (rotation, zooming, flipping, shearing, brightness) to enhance dataset robustness.
  3. Comparing model performance using accuracy scores, loss/accuracy graphs, and confusion matrices.
  4. Conducting EDA for dataset insights, including image distribution, quality, class imbalances, and sample visualization.
  5. Documening the process in a comprehensive README file.

Tools I'll use :- Keras , numpy , matplotlib , scikit-learn , tqdm etc.

Kindly assign this issue to me @sanjay-kv and lable it as level-3 if possible as I'm training and testing dataset for 5 different models which is resource and time consuming and also making doucumentation about which works better . I will also be doing EDA analysis for the dataset.

github-actions[bot] commented 4 weeks ago

Thank you for creating this issue! 🎉 We'll look into it as soon as possible. In the meantime, please make sure to provide all the necessary details and context. If you have any questions reach out to LinkedIn. Your contributions are highly appreciated! 😊

You can also check our CONTRIBUTING.md for guidelines on contributing to this project.

github-actions[bot] commented 10 hours ago

Hello @aaradhyasinghgaur! Your issue #221 has been closed. Thank you for your contribution!