[HardyLeung]

I think the OP mentioned that the snapshot is taken at the end of the exercise, basically to capture the statistics.

Even the image recognition part is not that difficult, assuming simple LCD display. The problem is whether they are able to handle the universe of exercise machines (or at least say 95% of exercise machines). One possibility is to ask users to submit image and if their machines are not found, ask for their help to take a few snapshots, etc. Given them a badge so they can brag about it (and even help spread the word).

In addition, I'm worried that the type of data provided is highly inconsistent amount machines. Can the OP easily combine incomplete information from different types of exercise machines?

If somehow the OP can manage this, I think it would be a great app, and yes, people are willing to pay money for this app.

[aspire007]

We got a strong image processing team and have figured out our core image processing algorithm. Our algo should be able to handle 80-90% of LCDs with ease.

If user is using a machine for first time, he has three options, (1) Select one of machines listed in app (2) User creates a machine profile (as shown in video link) (3) Search & download from server where we will continue to add new machine profile & also allow users to post their created machine profile