I usually avoid storing images in DB, because it leads to unnecessary large tables. Instead, I save them in a specific folder in the file system, and I save their references only in the database.
To illustrate, if I need to process images and save the original, a thumbnail of the original, and the image after processing, I implement the following algorithm.
1. Receive/get image, save in temporary folder.
2. Add entry in database table of image references.
3. Get primary key of the latter.
4. Create folder with a name based on primary key, e.g. directly
with the primary key itself (1,2,..).
5. Move image from temp folder to folder just created.
6. Create thumbnail and save in same folder.
7. Process image and save processed version in same folder.
8. Optionally, set a flag in DB which indicates that processing went through.
