Dumb Can Speak

Project Title

Dumb Can Speak

Project Area of Specialization Computer ScienceProject Summary

We have planned a technique for real-time hand gesture recognition and feature extraction using a webcam. The hand gesture is one of the methods used in sign language for non-verbal communication. It is mostly used by deaf & dumb people who faced hearing or speech problems to communicate with normal people. During this approach, we capture the image through a digital camera hooked up to the system. Initial the input image is preprocessed, and the threshold (separating an image into foreground values (black) and background values (white)) is used to get rid of noise from the image and smoothen (remove high spatial frequency noise from a digital image) the image. Once this applies region filling to fill holes within the gesture or the object of interest. This helps in the classification and recognition step. In this way, a dumb or deaf person will perform some specific hand gestures (alphabetic) and we will process those gestures and our system will display those letters in the form of words, and those words will speak loudly by the system.

•   The software will capture the hand gesture and convert it into words.
•    The software will read the converted words into voice.
•    Non-verbal to verbal communication
•    Reduce the efforts
•    Better communication between normal and abnormal people

Project Objectives

• Non-verbal to verbal communication
•    Reduce the efforts
•    Better communication between normal and abnormal people
• Dumb people can easily communicate with normal people. 

Project Implementation Method

• Image Acquisition: In this event, our system will capture images using the camera later these images are used to detect the hand gestures.

• Hand Detection: In hand detection, as we have taken the image, that image would contain many other objects. But our purpose is to detect the specific object which is a hand.
• Crop Hand: After detecting the target object next is to separate the hand from other objects, by simply cropping the hand.

• Preprocessing: The aim of pre-processing is an improvement of the image data that suppresses unwilling distortions or enhances some image features important for further processing, although geometric transformations of images (e.g. rotation, scaling, translation).

• Classification: Image classification refers to the task of extracting information classes from a multiband raster image. The resulting raster from image classification can be used to create thematic maps.

• Feature Extraction: After that, feature extraction techniques are applied to get features that will be useful in classifying and recognition images.

• Information to text: All the information gathered would be converted into alphabets and combining these alphabets generating a word.
• Text to Speech: Now these words would be converted into speech using some algorithms

Benefits of the Project

• Usability

      Our system will be easy to use and provide a user-friendly environment where a

      the person can easily interact with it.

• Efficient

      The system is efficient because it captures the image from the camera as input and after performing actions it detects the right gestures for sign language and converts it into speech. Our system takes short time to detect the right object and  

      provide maximum accuracy.

• Easy to deliver the message

      It is easy for the dumb and deaf person to deliver a message effectively and to

      the person who does not know the basics of sign language as our system will

      change it into voice.

Technical Details of Final Deliverable

The final deliverable project will be Software. And our software will produce an .exe file and can run on every desktop if the desktop has a window above 8. A sensor camera and GPU is required to run this software.

Final Deliverable of the Project Software SystemCore Industry ITOther Industries Telecommunication Core Technology Artificial Intelligence(AI)Other Technologies Shared EconomySustainable Development Goals Peace and Justice Strong InstitutionsRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	30
GPU GTX 1050	Equipment	1	24	24
Sensor Camera	Equipment	1	6	6