---
license: apache-2.0
task_categories:
- image-classification
language:
- mr
tags:
- marathi-sign-language
pretty_name: Marathi
size_categories:
- 10K<n<100K
dataset_info:
  features:
  - name: image
    dtype: image
  - name: label
    dtype:
      class_label:
        names:
          '0': अ
          '1': आ
          '2': इ
          '3': ई
          '4': उ
          '5': ऊ
          '6': ए
          '7': ऐ
          '8': ओ
          '9': औ
          '10': क
          '11': क्ष
          '12': ख
          '13': ग
          '14': घ
          '15': च
          '16': छ
          '17': ज
          '18': ज्ञ
          '19': झ
          '20': ट
          '21': ठ
          '22': ड
          '23': ढ
          '24': ण
          '25': त
          '26': थ
          '27': द
          '28': ध
          '29': न
          '30': प
          '31': फ
          '32': ब
          '33': भ
          '34': म
          '35': य
          '36': र
          '37': ल
          '38': ळ
          '39': व
          '40': श
          '41': स
          '42': ह
  splits:
  - name: train
    num_bytes: 279345185.402
    num_examples: 50099
  download_size: 241150104
  dataset_size: 279345185.402
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---
# Marathi Sign Language Detection Dataset Card

## Dataset Description
The Marathi Sign Language Dataset is a comprehensive collection of images designed to facilitate the development and training of machine learning models for recognizing Marathi sign language gestures. This dataset includes 43 distinct classes, each representing a unique character in the Marathi sign language alphabet. With approximately 1.2k images per class, the dataset totals over 51k images, all uniformly sized at 128x128 pixels.

## Dataset Summary
- **Total Number of Classes:** 43
- **Total Number of Images:** 51,600
- **Images per Class:** Approximately 1,200
- **Image Dimensions:** 128x128 pixels
- **Format:** Images (JPEG/PNG)

## Dataset Structure
The dataset is organized into 43 folders, each named according to the Marathi sign language character it represents. Within each folder, approximately 1,200 images are stored, each capturing a variation of the sign language gesture for that character.

```
/Marathi_Sign_Language_Dataset
    /अ
      0.jpg
      1.jpg
      2.jpg
      ...
      1999.jpg
    /आ
      0.jpg
      1.jpg
      2.jpg
      ...
      1999.jpg
    ...
    /ह
      0.jpg
      1.jpg
      2.jpg
      ...
      1999.jpg
```

## Data Collection
The images were collected through a controlled environment to ensure consistency in lighting, background, and gesture representation. Each image captures a clear and centered view of the sign language gesture, making it suitable for training robust machine learning models.

## Intended Use
This dataset is intended for researchers and developers working on sign language recognition systems, particularly for the Marathi language. It can be used to train, validate, and test machine learning models, including convolutional neural networks (CNNs), for gesture recognition tasks.

## Ethical Considerations
- **Privacy:** The dataset does not contain any personally identifiable information.
- **Bias:** Efforts were made to include diverse hand gestures from multiple individuals to reduce bias. However, users should be aware of potential biases and consider further augmentation or diversification if needed.
- **License:** The dataset is released under an open license, allowing for free use in academic and commercial projects with proper attribution.

## Access and Citation
To access the Marathi Sign Language Detection Dataset, please contact the dataset maintainers or visit the official repository. When using this dataset in your research or projects, please cite the dataset appropriately to acknowledge its contributors.

## Contact
For any inquiries or further information about the dataset, please contact the dataset maintainers at [vinayhajare2004@gmail.com](mailto:vinayhajare2004@gmail.com).