But is there a scientific basis to prefer this bet to, say, death-bed religious conversion?

As a bet, its possibility of payoff is consistent with the rest of what we know about the world, as opposed to religion, which is inconsistent. Furthermore, there are cases where things similar to this were done successfully, so extrapolation of technology might lead one to believe that we may be sufficiently good at reconstruction in the future. In comparison, no prayer has ever even slightly worked.

> "its possibility of payoff is consistent with the rest of what we know about the world"

What we know about the rest of the world is that this does not work, it will cause extensive damage even if it could work, so it would not necessarily re-generate the same person even if it did work [1] and that humans are notoriously awful at projecting technological advances 50 years out, to say nothing of our track record of projections even further.

Extrapolation of current technology, from any particular point in our history, points us in wildly incorrect directions. See: the history of people proclaiming us on the verge of vanquishing death. "too cheap to meter" energy. Nuclear cars. Flying cars. Lab-grown meat. Off-world colonies in the far-off year 2000. Choose an AI prediction. Choose a cloning prediction.

The notion that some technology has worked in the past, so a particular technology has some chance of working, is as flawed as "it appeared that my last prayer was answered, so my next one has a chance". Both require selective reporting from the available data to even entertain.

[1] There are no shortage of case studies of people who've sustained comparatively minor brain trauma and been reduced to mere shadows of themselves, if not manifested new and almost entirely different personalities.