Scribe

Scribe

Scribe allows one to create videos and live stream what is being written on any surface using a LiDAR to track what is being written, and where. Scribe can be applied to any surface, whiteboard, or blackboard, and make them connected devices

Motivation

You're sick, you can't leave bed, and you're about to miss class. You're relegated to begging your classmates for notes, and hoping you can get an accurate representation of lecture. With scribe, a professor will be able to live stream what is being written on the board during her class to her pupils, without the need to setup video cameras, special white boards, projectors and tablets, or any other system. By hooking up scribe, the professor enables remote viewing of exactly what's written on the board during the lecutre as she writes it.

Competitors

Smartboards are related to our project as they also seek to revolutionize the 'board' in a classroom. Our produt provides less functionality than a smartboard, but should be cheaper, and should be able to be installed anywhere, as well as provide our unique video stream featureset.

Cameras provide similar functionality to our product. We aim to change the way lectures are recorded by giving students an uninhibited view of whats being written on boards, and a clearer focus on the boards in order to provide a more accurate record of lecture.

Tablets conjoined with projectors also provide a similar functionality to our product, but storing live lectures as videos and recording them is unintuitive. Also, tablets don't provide the 'comfort' of lecturing on a blackboard, we seek to be as unintrusive in the classroom as possible.

Khan Academy videos and the main result of our product will actually be very similar. The main difference is rather than dedicate time alone with a computer to create a video lecture, our product will allow professors do this on the fly by creating the same style product during a live lecture.

Requirements

Shall

  • Accurately track writing on a blackboard by one user
  • Accurately generate an animation from sensor data
  • Have the ability for any battery powered components to last at least one lecture per charge
  • Be able to tell the difference between chalk and erasers touching the board *
  • Be easy for lecturers to install and use
  • Have final results accessible via internet, not just stored locally

Shall Not

  • Fall off the board while in use, or become disconnected due to board sliding.
  • Be obtrusive and get in the way of a lecturer
  • Shut off at any time during a lecture

Should

  • Have a user-friendly and secure web app to access the video
  • Be able to do this tracking/recording fast enough to keep up with a lecture in real time (for livestreaming purposes)
  • Be able to be set up on an arbitrary board
  • Be minimally invasive to the classroom
  • Have a reasonable solution if the connection is dropped briefly during use
  • Have final results viewable by any students registered to a course

Should Not

  • Be ‘laggy’ or annoying to use *
  • Be less convenient than sitting in a classroom to watch a lecture
  • Be time intensive to use, taxing on students or instructors
  • Cost over $400 to reproduce for use in a classroom setting *

* = Not Currently Implemented or Achieved

Components

Our project was a simple setup of:

Intel Edison with Custom Shield

Out-Of-Box Edison

Arduino Mega with Custom Shield

Out-Of-Box Mega

RPLidar

Spins at 5hz

Custom Shield

Wires lidar, Edison, and Mega properly

Software Description

On the software end, the arduino and the mega are simply running C code to interface with the LiDAR and send curl rquests to our server. The server is written in Node.js, it recieves the post requests and then sends the data along sockets to clients, who are also written with Node.js. The client then displays the images created on a canvas.

The server simultainiously saves data to a text file, in the form of x,y coordinates. These text files are saved so that a python script on the server, using matplotlib can create videos from them when lecures are over. These files are also able to be downloaded via clients.

Architecture

The device has a simple architecture outlined in the diagram below

Interaction Diagram

The interaction is simple, scribe is connected to a board, and sends data to a server, which sends it to clients

Use Cases

In its current iteration, scribe has limited use. It can however accomplish its central goal of recreating what is drawn on a board, so one could use it to record whats being written on a board and create videos or live stream it to a wide audience.

Video

18549 Team 2 from Nitsan Shai on Vimeo.

Team

Dillon Lareau

Gabe Hobeika

Jane Threefoot

Nitsan Shai