Blog Review: A view into Disney Research
GRAND PhD student Tiffany Inglis participated in a demo session attended by Joe Marks, Vice President, Disney Research. Later in the day, she heard Joe's presentation, part of the University of Waterloo's Distinguished Lecture Series. Here is her review.
Posted by GRAND NCE, April 4, 2011
GRAND PhD student Tiffany Inglis participated in a demo session attended by Joe Marks, Vice President, Disney Research. Later in the day, she heard Joe's presentation, part of the University of Waterloo's Distinguished Lecture Series. Here is her review. 
It was busy in the Computer Graphics Lab on March 24, the day of Joe Mark's visit. Joe is the Vice President of Disney Research, and he was visiting the University of Waterloo to learn about what type of research is done here and to share the ongoing projects at Disney. 

At 10:00 in the morning and all my fellow grad students were in the lab, eager to give Joe a 5-minute elevator pitch on their research. When Joe finally arrived, he toured the lab and spoke to everyone about their research, which ranged from comic creation and tools for animation to graffiti styles and texture synthesis. He seemed genuinely interested in our work and often found some way to relate it to what Disney was doing.

After lunch, there was a meeting set up for other grad students to meet Joe. It was a good turnout, not just because of the free food. We gathered around a table and started asking Joe questions about himself and Disney. 

His life story turned out to be quite relatable---an immigrant student from Ireland, he got his degrees from Harvard then worked at a number of jobs. Only recently, he heard that Disney was looking to do its own research in computer graphics, social sciences, etc. and needed people to lead the research teams. It was too exciting an opportunity for Joe to pass up, and he has been working there ever since. 

He has never been this busy and devoted to his work. The diversity of Disney's research areas means more learning in breath rather in depth, as one would expect of a typical research position, but that is exactly what Joe loves. 

Some of the research topics he described seemed deceptively simple (I was able to describe them to my mom later on the phone) yet clever and creative. My favourite was a story about purchasing photos. At Disney theme parks, there are numerous photo booths where one can purchase photos taken during thrilling rides. They are not exactly cheap, and sales have plummeted since people started taking cell phone photos of their photos. So Disney decided to implement the Name Your Price scheme, with which tourists pay any amount to get a photo. With this scheme, people paid on average a dollar per photo, which resulted in zero profit for Disney.

A social scientist suggested a twist---in addition to naming a price, tourists were told that half of the proceeds would go to charity. Now people were faced with a moral dilemma. As a result, the cheapskates stopped paying (or not paying) for photos. The charity received money that they would not have otherwise, and even taking off 50%, Disney still made more money than had they charged the regular price. Another unanticipated bonus was that the Name Your Price program became a fun, family activity with parents asking kids what they should pay, probably trying to throw in a little education about capitalism and philanthropy.

Later that afternoon, Joe gave a talk as part of the University of Waterloo's Distinguished Lecture Series. The audience was a diverse group of grads, undergrads and staff members with all sorts of different backgrounds. He quicklyintroduced himself, and delved into all the projects Disney is currently investigating. Some of the more traditional CG (computer graphics) topics included smoke effects and hair rendering. For hair rendering, he showed us several clips from the recent Rapunzel movie "Tangled". They were behind-the-scenes footage which made them all the more exciting. 

As for the social science aspect, he described the problem of traffic control in theme parks. Basically, Disney wanted a way of controlling the traffic flow without disrupting the experience for the tourists, so the researchers devised these handheld GPS-like devices. They were handed out to select visitors in place of a regular map. Once in a while, messages such as "hotdogs are 50% off at Magic Mountain" popped up to see how tourists would react. This study was just another example of a win-win situation in which the visitors received perks while the researchers collected useful data for learning about how to incentivize people. A regular social scientist can only dream of the amount of data Disney is able to collect.

As a PhD student in the area of Aesthetic Visualization, I am interested in various rendering techniques for 2D images. Often, the biggest issue is how to explain to someone else the real-world application of the final product. 

Unbeknownst to Joe Marks, he addressed this issue when describing one of their projects. The inspiration came from observing caustics patterns from a glass of water. Someone thought, "what if I could design a glass so that the caustics pattern showed Mickey Mouse?"And that was what they did. Using small glass tiles, their individual caustics contribution was calculated and mapped to the target image. The result is a piece of glass that can be sold as a memento at a souvenir shop---a brilliant application of NPR (non-photorealistic rendering)!

Unsurprisingly, the talk was successful and it was clear from the questions asked and the number of people that lingered afterwards to talk to Joe that a lot of students were interested in knowing more about Disney Research and its internship program.

Joe Marks is the Vice President, Disney Research. He is also the Chair of GRAND's International Scientific Advisory Committee. Tiffany Inglis is a PhD student at the University of Waterloo. She works on AESTHVIS, one of the 34 projects within the GRAND network.