Face-Text-Recognition

Select what feature you want to use
Image Selection for Facial Characteristics
Generating facial characteristics for given image
Display of Facial Characteristics for each user
Generating Image Text Recognition result
Assigning labels with probabilities to an image

Inspiration

It is irritating to see things and not know them as well as have data in non digitized format. Hence we decided to create
something which could solve this problem once and for all. And what better platform to make it accessible to all than android.

What it does

The app allows you to take 2 different paths:
1) Analyze an image to get facial characteristics such as Age, Gender, Hair Color etc.
2) Analyze an image for textual recognition and get it in PDF format or get to know about significant words in the images.

How I built it

Android is used as a platform for all API I/O.
1) Analyzing an image to extract facial characteristics was developed using Microsoft Azure Face API.
->First get the subscription key and endpoint to access Microsoft Azure Face API.
->Created android front end for providing input image to Azure Face API Calls and displaying the facial characteristics of each individual face detected to user.
->Post call to the API with image as parameter gives us the facial characteristics in JSON Format.
2) Analyzing an image to assign a label to an image along with their probabilities and get the text from an image into PDF format that is digitalizing the data. All of these is done using the Google Cloud Vision API.
->Generate API Key on Google Cloud.
->Created android front end for providing input image to Google Cloud Vision API Calls and displaying the labels/tags with their respective probabilities.
->Post call to the API with image as parameter gives us the facial characteristics in JSON Format.

Challenges I ran into

-> Integrating Microsoft Azure Face API with android.
-> Getting the response from the API Call in the desired JSON Format.
-> Conversion from .txt to .pdf in android.
-> Displaying the probability of labels in Android.

Accomplishments that I'm proud of

-> Using Microsoft Azure Face API to solve basic recognition problem.
-> Integrating Microsoft Azure API calls with android application.

What I learned

-> JSON
-> Microsoft Azure Face API
-> Android
->Google Cloud Vision API

What's next for Face-Text-Recognition

-> Using facial Recognition for security
Outline : While a user sets up an account scan his image and store it in Local database. Now when a user enters an incorrect password capture an image of the current user using front camera compare it with the stored image and mail the image to the owner of device.
-> Allow to search any possible word in the text.

Built With

android
azure
face-api
google-cloud
java
microsoft
vision-api

Submitted to

HackNC 2017

Created by

I have built native android application using android studio and applied material design specifications.
To call Microsoft Azure Face API and Google Cloud Vision API, i worked with AsyncTask in the background thread.
To show the probabilities progress bars has been used.

Khantil Choksi
Software Design Engineer, Back-End Developer, and Microservices. Enthusiastic for building production ready containers.
I worked on the back end and helped in accessing the Microsoft Azure Face API. Also, contributed in module of Text recognition, to find out the most significant text on the object/document. Created module to generate PDF of the document/picture, that firstly used Google Cloud Vision API to recognize the text.

Daxkumar Amin
To start with I worked to understand the Cloud Vision API, Microsoft Vision API and find which one is better.
Following it I worked on Google Cloud Vision API.
Next we(pair programming) developed a JSON Parser to return the Most Significant words in the image which will be allowed for google search.
Following it we worked on string to text file to pdf file conversion which allows to digitize hand written data.
At last I worked on code to display search result for significant words if asked by user.

Vishrut Patel
To start off, I contributed with the application of detection techniques in form of a mobile application. Followed by this, I contributed in understanding of the API's, which was a team effort overall. Given my limited exposure to android, I decided to utilize this opportunity to look into android development. Even though it was overwhelming at first, I eventually came around it and contributed to some significant steps in the android part of the application, like resizing of bitmap image for clearer output, etc. Overall, my knowledge related to android development has definitely increased significantly.

Aditya Upadhyay

Updates

Khantil Choksi started this project — Nov 05, 2017 05:26 AM EST

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.