Detection ofs Signlanguageminorppt1.pptx

Ppt

1. Expand Intro Hook Explore Explain Apply Share Evaluate # Sign Language Detection Using Machine Learning Python Project

2. Expand Hook Explore Explain Apply Share Evaluate Intro Harsh Vardhan Singh Sisodia (20/BCS/183) Rishabh Shakya (20/BCS/171) Yash Anand (20/BCS/203) # Presented By:-

3. Expand Hook Explore Explain Apply Share Evaluate Intro Introduction Objective Scope Literature review Methodology Gesture Classification Challenges Faced Result Project Requirements # Content

4. Expand Hook Explore Explain Apply Share Evaluate Intro 1) American Sign Language (ASL) is crucial for communication among Deaf and Dumb (D&M) individuals who can't use spoken languages. It involves nonverbal gestures understood through vision, facilitating meaningful exchange. 2) The project's goal is to create a model recognizing Fingerspelling-based hand gestures, combining them to form complete words. Specific gestures are targeted for training, as shown in the provided image. # Introduction

5. Expand Hook Explore Explain Apply Share Evaluate Intro  To create a computer software and train a model using CNN which takes an image of hand gesture of American Sign Language and shows the output of the particular sign language in text format converts it into audio format. # Objective

6. This System will be Beneficial for Both Dumb/Deaf People and the People Who do not understands the Sign Language. They just need to do that with sign Language gestures and this system will identify what he/she is trying to say after identification it gives the output in the form of Text as well as Speech format. # Scope Expand Hook Explore Explain Apply Share Evaluate Intro

7. 1. Electromechanical devices provide precise hand data, but costly and unfriendly; glove-based approaches are expensive. 2. Vision-based methods use webcams for natural human-computer interaction, reducing costs, but face challenges like variability. 3. Challenges in vision-based hand detection include coping with hand appearance, skin color, viewpoints, scales, and camera speed. Data Acquisition : # Literature review Expand Hook Explore Explain Apply Share Evaluate Intro

8. Data Acquisition : Expand Hook Explore Explain Apply Share Evaluate Intro We have collected images of different signs of different angles for sign letter A to Z.

9. Hand detection is achieved through webcam-acquired images using the Media Pipe library for image processing. The process involves detecting the hand, obtaining the region of interest, cropping, converting to a gray image with OpenCV, applying Gaussian blur, and converting to a binary image using threshold methods. Key tools: Media Pipe and OpenCV. We have collected images of different signs of different angles for sign letter A to Z. Data pre-processing : Expand Hook Explore Explain Apply Share Evaluate Intro

10. Similarly we have collected images of sign of B (Training the data) :- Data pre-processing : Expand Hook Explore Explain Apply Share Evaluate Intro

11. Similarly we have collected images of sign of C (Training the data) :- Data pre-processing : Expand Hook Explore Explain Apply Share Evaluate Intro

12. # Methodology Expand Hook Explore Explain Apply Share Evaluate Intro

13. The MediaPipe Hand Landmarker task lets you detect the landmarks of the hands in an image. The task to localize key points of the hands and render visual effects over the hands. Mediapipe Library : # Methodology Expand Hook Explore Explain Apply Share Evaluate Intro

14. # Gesture Classification Expand Hook Explore Explain Apply Share Evaluate Intro Now it’s Recognizing the Gestures / Sign language alphabets

15. 1. Dataset Challenge: Initial struggle with finding suitable square images for CNN led to creating a custom dataset for the project. 2. Filter Selection: Experimentation with filters, including binary threshold and canny edge detection, concluded with choosing the gaussian blur filter. 3. Model Improvement: Overcame accuracy issues by enhancing input image size and refining the dataset in later training phases. # Challenges Faced Expand Hook Explore Explain Apply Share Evaluate Intro

16. Finally, we got  97%  Accuracy (with and without clean background and proper lightning conditions) through our method. And if the background is clear and there is good lightning condition then we got even  99%  accurate results # Result Expand Hook Explore Explain Apply Share Evaluate Intro

17. This is the output of A alphabet with 100% accuracy :- # Result Expand Hook Explore Explain Apply Share Evaluate Intro

18. This is the output of B alphabet with 100% accuracy :- # Result Expand Hook Explore Explain Apply Share Evaluate Intro

19. This is the output of C alphabet with 99.99% accuracy :- # Result Expand Hook Explore Explain Apply Share Evaluate Intro

20. Operating System: Windows 8 and Above IDE: VScode Programming Language: Python 3.9 5 Python libraries: OpenCV, NumPy, Keras,mediapipe,Tensorflow Software Requirement: # Project Requirements Expand Hook Explore Explain Apply Share Evaluate Intro Hardware Requirement: Webcam

21. # Thanks!