Tech Assistant For Blind Person ( नेत्र 2.0 )

With नेत्र 2.0 Lets Make the bright future for all blinds and Dream of No Darkness. Main motive Never stop the fight for sight.

Comment

👨‍💻Inspiration

As we all know from the current data studied there are 39 million blind people on the earth and they face issues in their day-to-day life, be it accessing any written script or even recognizing their known. For this problem I build an Tech Assistant for Blind Person which I gave name नेत्र 2.0.Now I want to tell you all why I choose this name नेत्र 2.0 which may also clear the inspiration behind the project . One day when I was scrolling google feeds I read somewhere that in moon sign based Vedic astrology , person with name Naitra are always ready to do more in life and A proper education is what they think necessary and top most priority for all due to this reason the word Naitra came in my mind and 2.0 because version one (1.0) of our eyes as we all know is given by god and We design this product for blind people as second version who's god gifted eyes is not supported So I want to motivate all blind persons with this name नेत्र 2.0

🛠️What it does

The proposed system helps them to recognize people (using face recognition technique), detect any obstacle in the path (thus providing an ease in their day-to-day life), and convert textual scripts into audio signals (text detection techniques) that are provided to the blind person with the help of a microphone.

💻How we built it

There are three task mainly which have to built :

1) Text Reading
2) Object Detection
3) Face Recognition

For text detection I use optical character recognition module Py tesseract which extract the text from given input image and then this image we gave input to Google text to speech convertor so that it convert the text into speech for blind person. For Object detection I use Mobile- Net-SSD an Pre trained Model with 91 Objects as COCO dataset when any object from the given dataset is detected it save the object name and give input to GTTS convertor to convert into audio signals . This dataset is mostly used for navigation on road because it include the traffic signal sign and vehicle information which is helpful in Navigation from one place to the another . Then we trained an model for face Recognition using tensor flow and the data include the faces of family members so that blind person able to recognize their knowns or if any unknown person is there then also able to know . I want to show the complete block diagram of the system so it would be more clear .

💢Challenges we ran into

May face challenges with accuracy because there is always work pending with any machine learning model accuracy so we update our system models which gave best accuracy and also with the camera quality here I use less mega pixel camera but if We use high mega pixels then it gave accurate image as an input to the models .

🏆Accomplishments that we're proud of

As we all know when this type of system is not exist then there is an braille system which helped blind person in reading but the main drawbacks with this system :

1)All the books which normal students can read we need to design all the content in braille letters.
2) We hire extra faculty or trainers for training the blind person with this system .
3) This process is really very time consuming because the training took time.

so with our system we try to reduce this time as much as possible and blind person also sit with normal students and gain equal education .

I also Wanted to highlight the main objective of the project which proud us :
1)Physically Disabled Children Education :

Smart glasses in one such solution which enables blind or visually challenged people to “read” images .

2)Quality of Education Delivery for Visually Impaired Students :

All the visually impaired children also able to go and study as normal children .

3)Provide Security to Visually impaired Person :

Using this Navigation system blind person able to identify multiple objects and faces of family members . And last one is the cost reduction which I observed by doing market survey of the product

💹Market Survey

1) Eyesynth-Smart glasses by (Marcelo Alegre (Spanish Company))

-It allows blind people to ‘feel the space’. -It is not verbal. -Costly. (€ 499 =र40,800) -Recognize object

2)Oton Glass by Japanese company (Keisuke Shimakage)

-The smart glasses are designed to help dyslexic to read. -It will convert text to audio -Not much useful for blind.

💸Our Cost

Raspberry pi - 4350rs. , VR Box design - 300rs. , R pi Camera - 300rs. , Earphones - depends ,

#TOTAL 5000/-

So this is the main accomplishment which I achieve with cost reduction.

💯What we learned

We learned lots of new technology while making this product and most important work for needy people if I have tendency to do something for visually impaired people then why not I gave my time to build something innovative .

🕶️What's next for Tech Assistant For Blind Person ( नेत्र 2.0 )

1)Scene Recognition:

We are trying to add scene Recognition in future which gave proper idea of the environment what is happening near to the surrounding area of the blind people.

2)Patient Security:

We try to add security with face Recognition technique as when blind person is alone at home and someone is at the door then live feed of the IPcam is attached with the blind person Cam so its easy to recognize if any unknown person is there at the door or any known so its easy to take decision about further actions.

3)Translate Multiple Language:

Now English language is supported by the system in future we try to build with multiple language.

4)More Powerful Algorithms for Detection:

As a said earlier there is always the update required to the models which gave us best and accurate results.

5)Handwriting Reading:

Also try to reading the hand written text in future . Main motive in future is to implement नेत्र 3.0 , नेत्र 4.0 , नेत्र 5.0 etc. with multiple features.

Built With

Share this project:

Updates