During the design and planning stage of my puzzle game, in order to have something that would stand out from the crowd, I decided to include a novel user-interaction game-play that required mundane pattern algorithm as part of the puzzle core. Thinking about it then, it seemed like a very simple feasible concept.

But as I started coding the Image processing algorithms parts - all hell was let loose.

Some of the top end of the non-Image-Processing algorithms I've had to write can be described as monstrously difficult. But the difficulty level of the Image processing algorithms is in a super class of its own, and I'ill describe this as a mysterious beast, at least for me without a formal training in the field.

At the outset as a lonewolf, I estimated that I would finish the game in roughly 5 months. Its 11+ months now and still counting. And the Image processing problems are not getting easier. I'm even thinking of canning the game and starting a new development all over.

Lesson learnt/moral of the story:

[spoiler] Its a simple one, ....don't try to stand out from the crowd, join them [/spoiler]

Sounds like you've got a research project.

Games usually incorporate established research, even if they are relatively recent research papers. It is much easier to implement an existing algorithm or to adapt an established algorithm than it is to come up with something new and unique.

Research tends to be a lot of work. There is good reason why most research is incremental, with small steps beyond what was done before. Usually once you are familiar with the field you can come up with minor improvements along the way. Major transformations are relatively rare.

Consider finishing up your research and publishing it, you might come up with some amazing results that you can publish and share with the world, opening up new interesting areas for others to take advantage of.

You are very accurate, its been a research development indeed, with me discovering how difficult it is to determine the exact values of pixel colours (same colour) in a consistent manner,- its a real mystery when no pattern seem to emerge. And at the edges the pixel-colour gets really really fuzzy ...

My conclusions is moving towards this:- reading colours(decoding) is based on ratios rather than specific values

What deceived me at the outset was because writing colours is extremely trivial. So I had thought reading(decoding user ink colour input) would also be trivial. How wrong i was!!! In fact reading is exactly the opposite- extremely difficult

And then it got more bizarre when i realized values of different colours overlap. Hence when you are filtering (picking) a specific colour, say medium green (based on range and average calculated values) , then mysteriously Blue , which has different range and average value gets filtered (picked). This kind of overlapping was a shock to my brain

Its the difficulty in determining a time scale for this kind of development that bothers me. If I could put a figure on the time it would take then I would be able to reset.

Computers have it very hard to do things the brain finds easy (and viceversa). In particular, pattern recognition in general is hell.

Also, don't forget the whole blue dress incident.