Eyedentify

The Device “Eyedentify” is devised especially for the visually impaired to recognize people with the help of a cap fitted with camera and image processing unit.

Eyedentify

The Device “Eyedentify” is devised especially for the visually impaired to recognize people with the help of a cap fitted with camera and image processing unit.
Reduced Inequalities SDG 10

#SDG In Focus

SDG 10 : Reduced Inequalities

Rhiaan: My neighbor, Mr. Didwania, has a 100% loss of vision. 

Every morning, I see his wife guiding him on a walk, whispering in his ear when my sister and I were around to say ‘hello.’ It constantly nagged me that Mr. Didwania had to be told who he was about to meet.

Hence, it spurred me to brainstorm some device to assist him and wear that on his walks. Upon discussing this with him, I realized that such a  device would greatly benefit him and his wife during social outings. 

Henceforth, I formulated the problem statement for this project.

Problem Statement

To design a device to enable people who are blind to recognize who is around them without them having to ask many questions or having someone else describe their surroundings every time.

Preliminary Constrains

● It must work in any language (English, Hindi, Gujarati)
● It cannot be overly conspicuous
● It needs to be portable, since it must be able to work during walks
● It must be usable without external help

Solution

An easily wearable device that recognizes common faces and basic surrounding  objects, and describes these to the wearer through audio. A camera could be used for a  video feed, which is run through a machine-learning program in an embedded  computer to detect the faces of known people. A headband or cap was chosen since it  is an attire suitable for walking. These could also possibly be integrated into glasses  that he already wears. A remote-like device with differently-textured buttons that when pressed, have  different functions such as detecting the distance to the nearest object (possibly using  ultrasonic sensors), recognizing faces when pointed, and alerting for assistance. 

Device Idea: The basic idea is a cap with a Raspberry Pi (RPI) 3B+ computer attached to it and a Raspberry Pi Camera at the front, Powered by a generic ‘powerbank’ in the user’s pocket.

The RPI would run a code in Python, which after recognizing a face from a database, would play a pre-recorded audio file with the associated name. Henceforth, I named it ‘EyeDentify.’ 

Device Working: When a person who is blind is wearing the face recognition system enabled cap, the modular computer system (Raspberry Pi) of the face recognition system receives an input video frame captured by the camera. The input video frame may comprise a person’s face/facial features in front of the camera. After that, the modular computer detects the facial features in the received input. Subsequently, the modular computer system compares the facial feature detected in the received input video frame with one or more historic facial images stored in the memory of the modular computer system. When the facial feature is found in the memory on the comparison, the modular computer system sends the audio file with the name of the person that corresponds to the matched face to the amplifier (HXJ8002) and the speaker  (general purpose, 8 Ohm). When the face/facial feature is not found in the memory on the comparison, the modular computer system sends a default audio file with a sound indicating no matches to the amplifier and the speaker. In this way, the face recognition system enabled cap assists the person who is blind to identify the person in front of them.

Project Contributors

Rhiaan Jhaveri

View detailed project documentation

This Project Has Been Developed By The Alumni Of Innovation School

Rhiaan: My neighbor, Mr. Didwania, has a 100% loss of vision. 

Every morning, I see his wife guiding him on a walk, whispering in his ear when my sister and I were around to say ‘hello.’ It constantly nagged me that Mr. Didwania had to be told who he was about to meet.

Hence, it spurred me to brainstorm some device to assist him and wear that on his walks. Upon discussing this with him, I realized that such a  device would greatly benefit him and his wife during social outings. 

Henceforth, I formulated the problem statement for this project.

Problem Statement

To design a device to enable people who are blind to recognize who is around them without them having to ask many questions or having someone else describe their surroundings every time.

Solution

An easily wearable device that recognizes common faces and basic surrounding  objects, and describes these to the wearer through audio. A camera could be used for a  video feed, which is run through a machine-learning program in an embedded  computer to detect the faces of known people. A headband or cap was chosen since it  is an attire suitable for walking. These could also possibly be integrated into glasses  that he already wears.

A remote-like device with differently-textured buttons that when pressed, have  different functions such as detecting the distance to the nearest object (possibly using  ultrasonic sensors), recognizing faces when pointed, and alerting for assistance. 

Device Idea: The basic idea is a cap with a Raspberry Pi (RPI) 3B+ computer attached to it and a Raspberry Pi Camera at the front, Powered by a generic ‘powerbank’ in the user’s pocket.

The RPI would run a code in Python, which after recognizing a face from a database, would play a pre-recorded audio file with the associated name. Henceforth, I named it ‘EyeDentify.’ 

Device Working: When a person who is blind is wearing the face recognition system enabled cap, the modular computer system (Raspberry Pi) of the face recognition system receives an input video frame captured by the camera. The input video frame may comprise a person’s face/facial features in front of the camera. After that, the modular computer detects the facial features in the received input. Subsequently, the modular computer system compares the facial feature detected in the received input video frame with one or more historic facial images stored in the memory of the modular computer system. When the facial feature is found in the memory on the comparison, the modular computer system sends the audio file with the name of the person that corresponds to the matched face to the amplifier (HXJ8002) and the speaker  (general purpose, 8 Ohm). When the face/facial feature is not found in the memory on the comparison, the modular computer system sends a default audio file with a sound indicating no matches to the amplifier and the speaker. In this way, the face recognition system enabled cap assists the person who is blind to identify the person in front of them.

Solution In Action

Glimses of the making

Related Projects

Bullseye is a workshop designed to help Primary School kids know more about bullying and how to get out of it through various games and activities.
SDG 4

#sdginfocus

SDG 4: Quality Education

SDG 4

#sdginfocus

SDG 4: Quality Education

The Device “BirdSense” is designed to prevents the birds from coming near your residence. especially in major cities.
sustainable cities and communities

#sdginfocus

SDG : 11 Sustainable Cities and Communities

sustainable cities and communities

#sdginfocus

SDG : 11 Sustainable Cities and Communities