Datasets
BdSL-D1500 (Bangla Sign Alphabet dataset.)
Contains Images of :
1. One-Handed Bangla Sign Alphabets (38 Signs)
2. Two-Handed Bangla Sign Alphabets (36 Signs)
3. 12 Bangla Sign Numerals.
4. One Gesture
5. Different Background.
6. Number of Participants: 20 persons.
7. Total Number of Images: 132061 images.
8. Total Number of Bangla Sign Alphabet Classes: 87
Link: [Will be available Soon]
BdSLHD-2300 (Hand Segmentation Dataset)
Contains Images of :
1. One-Hand Image and Manually created Mask
2. Two-Hand Image and Manually created Mask
3. Different Background.
4. Number of Participants: 20 persons.
5. Total Number of Images: 2300 images.
6. Total Number of Maks: 2300 images
Link: [Will be available Soon]
BdSLHD-4k+ (Hand Segmentation Dataset)
Contains Images of :
1. One-Hand Image and Mask
2. Two-Hand Image and Mask
3. Masks are created n hand region. Face and other skin area is counted as background. Hand region = TRUE; Other Skin area, Background = False,
4. Number of Participants: 30 persons.
5. Total Number of Images: 4270 images.
6. Total Number of Maks: 4270 images
Link: [Available Based on Request. Email me if you feel interested to use the dataset, email me.]
BdSL_Dataset-Color_Coded_Fingertips
Contains Images of :
1. One-Handed Bangla Sign Alphabets (37 Signs)
2.Color Coded Fingertips.
3. Fingertip color combination: Black -> Thumb, Red-> Index, Green -> Middle , Blue -> Ring, Yellow -> Pinky.
5. White Background.
6. Number of Participants: 5 persons.
7. Total Number of Images: 45,958 images.
Link: Kanchon Kanti Podder, Muhammad Abdul Kadir, and Muhammad E. H.Chowdhury, “BdSL_Dataset-Color_Coded_Fingertips.” Kaggle, 2021, doi: 10.34740/KAGGLE/DSV/2344653.