Week 3 writing exercise

The Future of Artificial Intelligence in Visual Effects

In Seymour’s Article about the future of artificial intelligence and deep learning tools (2018) he explains that he is certain that ‘Artificial Intelligence (AI) is set to change the way VFX is approached and produced.’ He then goes on giving an example for machine learning, where he uses the example of fluid simulations which are really slow but have the potential to run in real time if there is enough previous data which the computer can learn from.

The authors of Emotional image color transfer via deep learning (2018) also tested the power of deep learning. It is introduced as a possible advancement in the future of color editing methods.They concluded that the color of an input image can be enhanced to its target emotion ‘without any user interaction’.

In contrast to all of that, Seymour (2018) also points out that this is not really artificial intelligence yet, the software is still coded by people and the programs are ‘innovative in their conception, but they are not indicative of general computer Intelligence’.

Seymour, CM. (2018) ‘A.I.’ for VFX at SIGGRAPH Part 1. Available at: https://www.fxguide.com/featured/a-i-for-vfx-at-siggraph-part-1/ (Accessed: 30/09/2018)

Liu, D., Liu, S., Jiang, Y. and Pei, M. (2018) ‘Emotional imagecolor transfer via deep learning’, Science Direct, (110) , pp 16-22, Available at: https://www-sciencedirect-com.ezproxy.uwl.ac.uk/science/article/pii/S0167865518300941 (Accessed: 02/10/2018)

Proposal

What are the last few per cent to sell the illusion?

Visual Effects and its road to photorealism

In this modern world, visual effects companies nowadays mostly strive for realism. Realism can be achieved by doing an extraordinary job in all the different departments such as lightning, rendering, texture, etc.., but what makes something computer generated truly believable? Which details are crucial for perfectly blending the real world with CGI (Computer generated images)?

The purpose of this research is to look beyond purely technical aspects of creating photorealistic CGI, such as algorithms for the perfect lighting conditions for example, but rather consider fine, sometimes invisible details that sell the effect. Things that go unnoticed by the naked eye but subconsciously give people a good feeling. If that special something is missing sometimes a viewer can get a feeling of something is not quite right but he or she might not able to really place what it is.

Research for this topic will include the uncanny valley, blending of live-action footage and CGI, human anatomy and how it is used in films like Ex Machina (2014) and the general question of photorealism itself. As these are all huge topics on its own, the outcome of this research will only include specific parts of each subject. Combining research from what seems to be a wide range of different areas is essential to answer a complex question such as this one. Fitting everything in 3500 words will be a challenge but is achievable with the appropriate amount of planning.

Credible sources will include looking at various types of books from the University of West London library, journals and online magazines, all linked to the previously mentioned content. Even Ted talks could be considered.

The findings of this research will on the one hand hopefully lead to a better understanding of what makes a visual effect, that is physically impossible in the real world, so good that the viewer doesn’t even start to consider it to be fake while watching. The distinction between a solid effect and one that makes people forget that what they are seeing cannot possibly be true. On the other hand, it could also potentially lead towards non-fiction effects that are so well hidden or well integrated that no one would even recognise it as computer-generated unless pointed out. Additionally, it should provide theoretical guidance to any future practical projects involving the blending of live-action footage with CGI, or photorealism overall.

In general, it could be said that the goal of this research is to understand what it takes to well and truly fool people’s minds with computer-generated content. The small details that make the difference.

Reference List:

Ex Machina (2014) Directed by Alex Garland [Film]. Universal City, Calif: Universal Pictures

Structure

Final Essay

This is a link to my final essay. The layout has been a bit corrupted during the process of uploading a pdf to a google doc.

https://docs.google.com/document/d/1Yp40U4tj0VVsO4jkr3iSSqQBOV6r_cwXhEgjPuDrQS4/edit?usp=sharing