ahmedfaiyaz/OkkhorDiffusion-Ekush

Okkhor Diffusion

Okkhor Diffusion is a category of Denoising Diffusion Probabilistic Models designed to generate images of Bangla handwritten characters. This model card corresponds to Okkhor-Diffusion trained on Ekush dataset.

Variants

• Okkhor Diffusion trained on Banglalekha-Isolated
• Okkhor Diffusion trained on CMATERdb
• Okkhor Diffusion trained on Ekush

Usage

from diffusers import DiffusionPipeline
import torch
device="cuda"
pipeline = DiffusionPipeline.from_pretrained(
              "ahmedfaiyaz/OkkhorDiffusion-Ekush",
              custom_pipeline="ahmedfaiyaz/OkkhorDiffusion",
              embedding=torch.int16
            )
pipeline.to(device)
pipeline.embedding=torch.tensor([10],device=device) # 'ও': 10
pipeline(batch_size=1,num_inference_steps=1000).images[0]

Mappings

OkkhorDiffusion-Ekush
Character	Serial
অ	1
আ	2
ই	3
ঈ	4
উ	5
ঊ	6
ঋ	7
এ	8
ঐ	9
ও	10
ঔ	11
ক	12
খ	13
গ	14
ঘ	15
ঙ	16
চ	17
ছ	18
জ	19
ঝ	20
ঞ	21
ট	22
ঠ	23
ড	24
ঢ	25
ণ	26
ত	27
থ	28
দ	29
ধ	30
ন	31
প	32
ফ	33
ব	34
ভ	35
ম	36
য	37
র	38
ল	39
শ	40
ষ	41
স	42
হ	43
ড়	44
ঢ়	45
য়	46
ৎ	47
ং	48
ঃ	49
ঁ	50
ব্দ	51
ঙ্গ	52
স্ক	53
স্ফ	54
চ্ছ	55
স্থ	56
ক্ত	57
স্ন	58
ষ্ণ	59
ম্প	60
প্ত	61
ম্ব	62
ত্থ	63
দ্ভ	64
ষ্ঠ	65
ল্প	66
ষ্প	67
ন্দ	68
ন্ধ	69
স্ম	70
ণ্ঠ	71
স্ত	72
ষ্ট	73
ন্ম	74
ত্ত	75
ঙ্খ	76
ত্ন	77
ন্ড	78
জ্ঞ	79
ড্ড	80
ক্ষ	81
দ্ব	82
চ্চ	83
ক্র	84
দ্দ	85
জ্জ	86
ক্ক	87
ন্ত	88
ক্ট	89
ঞ্চ	90
ট্ট	91
শ্চ	92
ক্স	93
জ্ব	94
ঞ্জ	95
দ্ধ	96
ন্ন	97
ঘ্ন	98
ক্ল	99
হ্ন	100
০	101
১	102
২	103
৩	104
৪	105
৫	106
৬	107
৭	108
৮	109
৯	110

Citation

@ARTICLE{10445466,

  author={Fuad, Md Mubtasim and Faiyaz, A. and Arnob, Noor Mairukh Khan and Mridha, M.F. and Saha, Aloke Kumar and Aung, Zeyar},

  journal={IEEE Access}, 

  title={Okkhor-Diffusion: Class Guided Generation of Bangla Isolated Handwritten Characters using Denoising Diffusion Probabilistic Model (DDPM)}, 

  year={2024},

  volume={},

  number={},

  pages={1-1},

  abstract={Bangla has a unique script with a complex set of characters, making it a fascinating subject of study for linguists and cultural enthusiasts. Unique in some of its similar characters which are only distinguishable by subtle differences in their shapes and diacritics, there has been a notable increase in research on Bangla character recognition and classification using machine learning-based approaches. However, Handwritten Bangla Character Recognition (HBCR) training requires an adequate amount of data from a diversely distributed dataset. Making diverse datasets for HBCR training is a challenging and tedious task to carry out. Yet, there is limited research on the automatic generation of handwritten Bangla characters. Motivated by this open area of research, this paper proposes a novel approach ’Okkhor-Diffusion’ for class-guided generation of Bangla isolated handwritten characters using a novel Denoising Diffusion Probabilistic Model (DDPM). No prior research has used DDPM for this purpose, making the proposed approach novel. The DDPM is a generative model that uses a diffusion process to transform noise-corrupted data into diverse samples; despite being trained on a small training set. In our experiments, StyleGAN2-ADA had notably inferior performance compared to Okkhor-Diffusion in generating realistic isolated handwritten Bangla characters. Experimental results on the BanglaLekha-Isolated dataset demonstrate that the proposed Okkhor-Diffusion model generates realistic isolated handwritten Bangla characters, with a mean Multi-Scale Structural Similarity Index Measure (MS-SSIM) score of 0.178 compared to 0.177 for the real samples. The Fréchet Inception Distance (FID) score for the synthetic handwritten Bangla characters is 5.426. Finally, the newly proposed Bangla Character Aware Fréchet Inception Distance (BCAFID) score of the proposed Okkhor-Diffusion model is 10.388.},

  keywords={Deep learning;Handwritten character generation;Generative Model;Denoising Diffusion Probabilistic Model},

  doi={10.1109/ACCESS.2024.3370674},

  ISSN={2169-3536},

  month={},}