Draw Your Face in a Movie Poster Robot

Our group has had an incredible journey piecing together our robot together over the last 6 weeks. After building the CNC machine from scratch, we had to develop an idea that changed the nature of the CNC and used it in a completely new way than what was intended by it’s creators.

Concept: Create a robot that take a picture of the user, and then draw that user’s face as a character on a movie poster.

 

Our team had various design challenges we had to overcome. Some of the considerations we had to think about included:

  • Designing a method by which to hold a writing utensil. This is important as we needed the unit to flex just enough to provide a natural artistic stroke to the robotic gantry.
  • Designing the CNC bed to provide ample support, pressure and texture to the pen from above.
  • Determining the proper writing utensil
  • Converting an image from it’s original format to SVG, making it Black and White, then editing it so that the image is only composed of outlines.
  • Making sure the camera recognizes the user’s face
  • Determining where to place the cropped and converted image of a user’s face into the appropriate area on the movie poster

 

Method to Hold the Pen

We experimented with various types of ways to hold the pen using the original router holder. We ended up designing our own plate and 3D printing our pen holders for maximum flexibility and precision.

 

 

Designing the CNC Bed

We wanted to provide the proper surface for the pen to write on the bed. We created a bed out of acrylic, to make sure that the holes were covered. We then placed a thin sheet of foam and thicker poster paper on top to provide a firm, cushy and smooth surface to put paper on so the pen would easily glide while drawing.

 

Deciding the Proper Writing Utensil

We tested with various types of pens and markers. The way our machine is set up, it made the marker lines way too thick, so we opted to go with a ball-point pen, as it easily glides along the paper.

 

 

Computer Vision, Image Conversion and Process Flow

Our process flow is the following:

 

Select movie poster. Title it ‘poster.jpg’

Take image of user

 

Python script detects the user face on the movie posters

Select the face you want to swap with and swap faces

 

Convert File to Black and White & SVG