Author

• Name: Oladayo A. Luke
• LinkedIn: https://www.linkedin.com/in/oladayo-l-2414a675/
• Publication: https://www.proquest.com/docview/3054372234

Created Date

• Date: May 09, 2024

Dissertation Source Code

This repository contains the dataset and source code developed as part of my dissertation research on enhancing sign language recognition and hand gesture detection using Convolutional Neural Networks (CNNs) and data augmentation techniques. https://www.kaggle.com/datasets/oladayoluke/enhanced-sign-language-mnist-dataset/code

Dataset

The Sign Language MNIST dataset and the Enhanced Sign Language MNIST dataset used in this research can be downloaded from their respective sources. Please refer to the dataset link below for more information on how to download and use the datasets. https://www.kaggle.com/datasets/oladayoluke/enhanced-sign-language-mnist-dataset/

k_train.csv: contains the base train samples.

k_test.csv: contains the base test samples.

g_train.csv: contains the generated train samples.

g_test.csv: contains the generated test samples.

Enhanced Dataset Classes and Count
Label	Description	Count
0	A	2900
1	B	2885
2	C	2897
3	D	2884
4	E	2899
5	F	2894
6	G	2882
7	H	2892
8	I	2893
10	K	2888
11	L	2893
12	M	2893
13	N	2885
14	O	2885
15	P	2877
16	Q	2877
17	R	2881
18	S	2888
19	T	2877
20	U	2870
21	V	2871
22	W	2874
23	X	2874
24	Y	2893

## Project Structure.
Phase 1
	Stage 1 - 24 : Data generation
Phase 2
	Stage 1 : Hyperparameter search
	Stage 2 : Data Augmemtation methods search
Phase 3
	Stage 1 : DAM1 on all CNN models
	Stage 2 : DAM2 on all CNN models
	Stage 3 : DAM3 on all CNN models
	Stage 4 : DAM4 on all CNN models
	Stage 5 : DAM1 on all CNN models
	Stage 6 : DAM2 on all CNN models
	Stage 7 : DAM3 on all CNN models
	Stage 8 : DAM4 on all CNN models
	Stage 9 : Examination of Data

Citation

If you use this code or the datasets in your research, please cite the following dissertation:

Oladayo Luke. (2024). Enhancing Sign Language Recognition and Hand Gesture Detection using Convolutional Neural Networks and Data Augmentation Techniques. (Doctoral dissertation, Nova Southeastern University). https://www.proquest.com/docview/3054372234

Closing Remarks

Thank you for exploring the source code for my dissertation research. I hope this code provides a useful starting point for your own research in sign language recognition and gesture detection. If you have any questions or feedback, please feel free to contact me through my linkedIn.