Storing images in database rather than storing their path (on disk) is mostly avoided due to performance perspectives. The obvious benefit is that images are also backed up along with database and there is no need to take care of images backups separately.
On the other hand utilizing images path and getting images from disk storage system is optimal. However these are design considerations that appear rare as compared to take over of existing database.
Some examples of these limitations are
- Images may not be included in distinct list
- Images can not be used while using UNION, however you may use these with UNION ALL
- Len() function may not be used with images, you may use DataLength() to get bytes used by image
If insert and update operations for images are rare in your scenario then it would be better option to store images in binary format when inserted along with inserting path for image on disk. Later image may be retrieved from disk through path stored in database. It would make sure that images are also backed up along with database.