将多张图像合二为一的算法

Question

I'm looking for an algorithm that will merge multiple images into one image. The type of solution I'm looking for would not skew aspect ratios and would cause minimal overlap, but would lay images into a frame of a given size.

Example, say I have 6 images of carpentry tools and I want my algorithm to lay them all out in one frame 300px by 300px. Images can be scaled if aspect ratios are maintained, and you want all images to be visible with minimal overlap. No rotations either.

Answer 1

The problem you are solving is np hard like bin packing but you can use heuristic algorithms to achieve that in reasonable time. I would suggest Genetic algorithms for this case using fitness function which is proportional to empty space and inversely proportional to square error in sizes of the images.