I consider the problem of reducing storage space when storing images of several JPEGs together as one large image. The main intuition is that images tend to have some similarities (for example, those that were taken in one place or at the same time), and can we use this similarity to save space?
General stream: JPGImage Input → Each image converted to RGBImage Palisas → Reorganize similar fragments RGBtogether → Convert to format again JPG. Naturally, when extracting images, we need a reverse process.
Using the DC factor of the Y component as a similarity measure for tile reorganization, I got ~ 8% space savings for 10 images. When I do this for 100 images, the savings are reduced to ~ 3%.
How to save after a tile reorganization - that is, what part of the JPEG coding process uses this image tile reorganization?
Instead of the Y-component DC coefficient, there are some other metrics that you might think would be better to use JPEG encoding
Corrected version:
Is there an image format other than JPG that can better use similar similarity when combining multiple images? For example, how is PNG?
source
share