How to remove background elements (people, tripods) in a moving 4K video?

[Marvolo]

Dear all,

I'm currently in the process of editing a wedding film for a close family member. I wouldn't consider myself a professional video editor, at least it's safe to say I used to be a lot more familiar with all this stuff some 10-15 years ago when I used to do it more regularly than nowadays. Also, I'm currently using DaVinci Resolve Studio, which is by far a much different work flow than when I used Adobe Premiere or Magix Video Deluxe back in the days.
Despite these incapabilities, I'm trying to make it as professional as possible. After all, it's not just a random video clip but a wedding film, intended to be re-watched at anniversaries or whenever the mood is right.

Now I'm running into the first problems:
The event was filmed simultaneously with three different cameras, so as to be able to switch between different shots (Shot & Reverse Shot etc.). Problem now is that in the reverse-shots footage, the person filming the regular shot can be seen filming in the background as well as a tripod containing a stationary third camera.

Ideally, I would like to remove the filming person as well as the stationary tripod camera from all the reverse shoot footage but so far I haven't found any convincing way to achieve this goal. DaVinci Resolve Studio may have an implemented feature to remove background things and I tried it of course, but it doesn't seem to be coping well with my footage. The difficulty could lie in the fact that the filming person that is to be removed from the footage is sometimes close or right behind another person in the foreground which can't be removed. So the two are overlapping each other sometimes.

I was thinking if maybe current, up-to-date, state-of-the-art AI tools might be able to achieve what I want in a flick of the wrist, yet I do not feel very comfortable uploading de-anonymized footage to a cloud based AI so it can work, learn from and use the people displayed in the footage. If anything, it would have to be some sort of AI working 100% offline and local on my machine.

I have uploaded a short sample clip and blurred out all the faces for anonymity purposes. What do you think of it? Easily achievable or perhaps better to leave it in the background as it is?

Thanks to everyone trying to help!

PS: on another, un-related topic: it's absolute hell trying to use footage from 3 different cameras in one project. The stationary tripod Panasonic camera was the oldest in terms of technological features and filming in 1080p 50fps (constant frame rate) and in 4:2:0.
The person filming in the background is using a SONY Z90 4K camcorder filming in 4K 25fps (constant frame rate) and in 4:2:0.
The reverse shots as seen in the uploaded clip were shot on an iPhone 16 Pro Max using 4K at 25fps (variable frame rate ranging from 23,810 fps to 26,316 fps) and in 4:2:2 ProRes. The colors of the iPhone footage look a lot brighter, more saturated and more alive than the ones of the Sony Z90 or Panasonic 1080p50.

Now try matching all of these resolution- and colorwise in one 4K timeline.... But I'll leave this for another thread maybe..