Inspiration
Growing up, my family had the annual tradition of visiting the habilitation centers for deprived and less well-off persons. These experiences engendered in us a deep appreciation of how lucky I had been and of the need to assist such individuals. In this project, we had decided to create a navigation device for visually challenged persons to assist them in movement, and to increase their independence and to assist them in their path to self-actualization. The sight of a smile on the faces of individuals who have long been deprived of such basic functions is a source of inspiration in itself.
What it does
This project guides the individual in their quest to move from one location to another. We have attempted to create software for a navigation device for visually challenged individuals, which uses speech-to-text technology. This technology is integrated with geo-mapping to provide a perfect path from the current location to the final destination. Our software also uses computer vision technologies including object detection and image processing to detect possible obstacles in the path and guide the individuals to their desired destination with uttermost safety.
How I built it
The process of making this project first involved importing SpeechRecognition, PyAudio modules, and geocoder. Then after we had initialized the Speech Recognition module, we created a Microphone object which would record the speech of the said person, dictating the location. Then, the speech is converted to text and the final destination is extracted from it. Following this, there is a route generated from the current location of the individual to the desired final destination geocoding and Google Maps. We used computer vision which included technologies such as object detection and object tracking. We also used Yolo algorithm for multiple object detection. This computer vision model could detect up to 80 different objects which include car, person, chair, table and many other objects s. The model also provides the probability of a given object.
Challenges I ran into
It was a tough challenge for all three of us to accomplish this task. We had never performed such intensive python programming but the challenge got us excited. Converting speech to text is a day to day feature which we use in mobile, laptop and other multiple other applications. Installing the correct modules to our system and ensuring that those modules run on our program occupied most of our time. Displaying a generated route through google maps was also an obstacle but it was cleared through our teamwork. Additionally, the main challenge that we faced was in the field of obstacle detection in the course of our path. It required computer vision technology and we were only beginning to take baby steps in the field. Our team leader Shivansh being a senior had sufficient knowledge of computer vision and pulled us through this challenge, making a program dealing with object detection. The main part of this project was to integrate all the three components and create a functioning prototype so that we can make additional improvements in the future to assist visually challenged individuals.
Accomplishments that I'm proud of
The joy that we could see on each other faces when we ran smaller parts of our program was unbelievable. Being friends for a year it was a memorable moment for us to witness this. For us, this will be our first program together and we are very proud of it. It is not the result that matters but the hard work that was put in.
What I learned
We learned new technologies like speech-to-text technology, geocoding and object detection technology. We also learned time management, teamwork, and cooperation as well as prioritizing under deadlines. It was a learning experience and we are all the richer due to the experience
What's next for TRIMURTI
we will be keeping our option open and implement the new feature as we go along. This being our first project is also very close to each other. We create a working model for this and run much efficiently. It will be seeking to mass produce for social good.
Log in or sign up for Devpost to join the conversation.