Can I write a paper on a method that is novel but yields similar results compared to existing methods?

I have developed a method that gives similar results (in few cases: better) compared to existing methods and I want to publish it. Can I do this and if yes, how do I motivate it? More specifically:

I am a undergraduate researcher, investigating methods to detect X in images. We have a solution using two images and another method proposed in the literature before also used two images.

Recently we discovered that we can solve the same problem in a very different manner if we use three images (taken differently) and still get similar results. This technique will be based on a particular observation and it is not discussed in the literature. Does it make a good paper given that the problem can be solved using two images with similar results?

How should I write and explain it in my paper about why I am using three images instead of two, because the results obtained are similar? The way in which the three images are taken is different from taking two images and our observation in the three-image-case seems to be novel.

Edit (from comment): In some cases our previously proposed method which uses 2 images gives inaccurate results compared to 3 image solution.

• It's hard to say anything without knowing more of the field. Are images easy to come by in this situation? Is capturing an image cheap (e.g. consumer products) or challenging (e.g. distant astronomy features). Is this a mostly theoretical paper where you're trying to just propose a novel idea or is it attempting to solve a more real-life issue? Regardless of the answers, this seems like a good question for your adviser or a more experienced researcher in your group.
– Ric
Dec 4, 2015 at 19:27
• 1). Capturing images is cheap. 2). Paper is not theoritical, implementation is done and results will be shown. 3). It is a real life problem while taking images. Dec 4, 2015 at 19:41
• Is there any advantage (such as higher accuracy, more confidence of results, ...) in using 3 instead of 2 images?
– mdd
Dec 4, 2015 at 19:53
• I believe this inaccuracy (even if it is only in some cases) is something that you could focus on in your new paper. In case your new method with 3 images can eliminate it in a substantial number of cases, I think you have a quite convincing paper.
– mdd
Dec 4, 2015 at 20:02
• Dec 5, 2015 at 6:36