A typical procedure would be as follows:
The film reels are mounted, one at a time, onto a "flying-spot scanner". A more-or-less conventional film-feed mechanism moves the film, one frame at a time, across a glass plate; a swiftly-moving spot of light scans the image, and the light which passes through is split into 3 primary colours [Red, Green, Blue] and picked up by photosensitive elements. At this stage, colour and black/white balance can be corrected, by changing the gain on the photosensors. The three primary-colour signals are, typically, then matrixed to produce the "YUV" signals needed for video, converted into a standard video signal, and recorded to video tape — ordinarily digital tape, and often high-definition. Sometimes the primary-colour signals are stored as "4k" RGB on hard drives, depending on how the film was scanned.
At this point, it is possible to take digital video and go through and apply different filters and techniques to it, with the intent of cleaning up scratches and so forth. Whatever else is done, the next step is typically to perform MPEG-II compression at Standard Definition resolution, and record the result onto DLT, a special kind of video tape which is matched to DVD in its properties. This tape is then given to the DVD Authoring people, who insert chapter stops and design menus and all that, and then the whole result is burned to a special "authoring" DVD-R and shipped off to the pressing plant for replication.
This is an extremely simplified view of the process, but perhaps it will tell you what you want to know.