Ameer Holmes Capstone - Important *MUST SEE*
Capstone: Cartoon Cam
The goal of my project was to make a working cartoon camera from scratch and I have succeeded. The project in terms of feasible goals for a year's worth of learning and doing for this capstone was indeed a success, though the project itself standalone is not yet done. This capstone was definitely beneficial to the completion of this project. However so, it’s functionality as a cartoon camera for me and my specific wants for specified functionality of a grander project; this more so helped me see the flaws in the way I was going out in completing this task.
The major flaw was in part my lack of knowledge on the subject, and taking a year to understand how coding worked for the needs of this project specifically was extremely beneficial. Without doing it the way I have done it so far, by translating an input image to an output image based on a set of rules I created, I wouldn’t have the skill required, nor the insight to recognise what the next step was. While currently code is a set of rules is basically a filter that during the process of making this I constantly experimented with and changed to get it as perfect as possible in most situations, depending on lighting (a hard one) and colors; I don’t think it will ever be perfect the way that I want it by examining the image myself and creating a set of rules based off of what I see and what I think will happen by adding or changing a value. The amount of arbitrary rules that I would have to solve would take years and not to mention the fact that I would need the knowledge base of an online code dictionary to get this done. This brings me to the conclusion of creating something in which the computer (having an infinite amount of online knowledge) would create the rules to.
The next step, or the more reasonable approach to a cartoon cam is to make a program that uses batch learning or machine learning. This would not be an easy task either, but it is a much less impossible task. The original vision for this project (not just the capstone) was to create a website where people could upload their own code to have a massive database of selectable cartoon cams for a physical cartoon cam viewer. The rout I am thinking about now, more feasible to myself and the average future consumer of such a project is to have code that takes on two videos. The first video takes in a real life scene, and the second video is an artist's translation of the scene. The code will examine the two and recognise the difference and then create a set of rules for how to translate video from there on out. Not only does this eliminate the need for the user interface to require users know how to code, but it also adds a level of calculation from artists and a level of calculation from a computer that I would never be able to achieve on my own. This capstone allowed me to recognize this for moving forward in continuation of the project. The evolution of this thought \was in my opinion the most successful part of the capstone.
Timeline of code for project:
https://docs.google.com/document/d/1mbFksdmPfIC2YdsT_rkrfCW58LXOGRpFzaLMoxAnXzw/edit?usp=sharing
Cool Pictures of The Three Finales