Arbab Tufail

Academic

Career

Verbatim text

Path Visualization 
 
An android app for visually impaired and blind people for the visualization of the path and notify 
the user about the detections in its way. This app utilized the device camera which detects the 
obstacles in its way like the potholes, vehicles, pedestrian, pole, fire hydrants etc. Furthermore, an estimated distance from the camera to the obstacles was also calculated with the help of the 
average height of the obstacles. This app uses Speech-To-Text and Text-to-Speech feature to 
communicate with user. Different datasets are merged and trained on Yolov3 model. The working 
of the app is that first the user would login to the app using his voice and the app would start a live 
feed using the camera and the results of the model will be displayed on the camera and focal length 
formula is used to calculate the distance using the average height of the objects in its way and the 
app would notify the user about the path. 
Features include: 
-Realtime detection of objects using Yolov3 Model and Django restframework API to pass upload 
results to the server and Volley multipart request to fetch the data from the server in the app. 
-Distance Estimation using focal Length formula. 
-App use Speech-to-Text and Text-to-Speech feature. 
 
 
 
 
 
 
 
 
 
 
Technology Used: 
Deep Learning (Yolov3), OpenCV, pytorch, TensorFlow,  
Keras, Python, Django Rest framework, PyCharm, 
 Android Studio, Postman, Git 
Supervisor Name: 
Mr. Saad Salman 
Group Members: 
Mohammad Wasique Sheikh (I17 - 0117) 
Arbab Tufail (I15 - 0136)

AI enrichment