iLIP

Here at Middlebury College, students use their college IDs on a daily basis, from swiping in to dining halls with their meal plan credentials to entering into dormitories across campus. However, these IDs only display a small amount of a student’s information; his/her name, birthdate, and ID number.

​

There are several scenarios where having an application that can return student user address could be useful. In cases where lost IDs need to be returned, or in situations involving Public Safety or student safety concerns, student IDs alone do not provide pertinent full information about where the student lives or what year the student is. Moreover, the mail center faces a challenge where incoming student mail often does not have the student’s box number and the box number needs to be verified or in some cases, the incoming student mail belongs to alumni.

 

We as Middlebury and Image Processing students thought we could solve this issue by creating an application that students, faculty, and public safety can use to take a picture of a recovered ID and receive information from Middlebury’s school-wide directory. Not only would we be able to use techniques of filtering and thresholding from the content of our course, but we would also be able to make a multi-purpose application to be used in different situations.

 

In the initial stages of our project, the two biggest sources of influence we researched were the pytesseract library and a general html scraper library; both critical libraries for our project, along with the many other libraries we picked up from coursework. The pytesseract Optical Character recognition would be the main method of translating the pixels on a given image to mutable string text. Additionally, the HTML scraper would be used to then pass the extracted string from the image into the go.middlebury.edu/directory/ website to then return information on a student. Along with these two main libraries, we also used the libraries scikit-image and Pillow for general image processing purpose, like reading in files or converting images to grayscale. Understanding these two libraries and the few other imported libraries we used helped our team visualize how our project would flow.

​

Click on the links below to view in-depth details about our application and work process.

​