BISINDO Indonesian Sign Language: Alphabet Image Data

The dataset comprises images of BISINDO letters from A to Z, captured using different devices, backgrounds, and individuals, with the variation in backgrounds posing challenges during data preprocessing to distinguish hands from real-world elements. Collected from volunteers who provided consent and ethical approval, the dataset is openly accessible to the research community. It includes two versions: original images, which are larger as they come directly from volunteers' devices, and compressed images that are preprocessed and smaller in size. This dataset, representing 26 letters, is suitable for multiclass classification tasks in computer vision, particularly in developing recognition models for Indonesian Sign Language (BISINDO). It provides essential resources for communities and institutions working with sign language to train machine learning models focused on sign language and gesture recognition. The dataset includes various image types—images with backgrounds, background-removed images, binary images, and encodings—allowing for the development of models adaptable to real-world conditions. Additionally, this dataset contributes to the expanding resources for further research in BISINDO and sign language recognition.