Recursive File Renaming & Natural Sorting (Bash)

NB! I have already started a thread on this subject on ComputerHope (with the thread title "Recursive File Renaming & Logical Sorting"). However, on ComputerHope they are perhaps more specialized in Windows Command Prompt, and not that much in shell scripts for Bash (I guess).

I have a bulk renaming problem that I have to tackle in some way with the help of a shell script for Bash. I should write a shell script to recursively rename all the image files in a folder and its subfolders to "001", "002", "003" and so forth. The aim is to preserve all the file extensions (and my image files have many different extensions: jpg, jpeg, png, bmp and so on). All the folder names should also remain unchanged. Here's a short example showing what I mean:

rbots-ctf\cf-1024.jpg ----> 001.jpg
    rbots-ctf\cf-1280.jpg ----> 002.jpg
    rbots-ctf\cf-1440.jpg ----> 003.jpg
    rbots-ctf\cf-1600.jpg ----> 004.jpg
    chrome\backblue.jpg ----> 001.jpg
        chrome\fade.jpg ----> 002.jpg
        chrome\skulltag.bmp ----> 003.bmp
        dmoz\newtip1.png ----> 001.png
            dmoz\newtip2.png ----> 002.png
            dmoz\newtip10.png ----> 003.png
            dmoz\newtip11.png ----> 004.png
            dmoz\pong.gif ----> 005.gif
            dmoz\pong.png ----> 006.png

...and so on!

As you can see, once the script is applying new filenames to the files that it can find, it starts from "001" in each folder.

Thus far, I have tried to put together a batch for Windows Command Prompt and a shell script for Bash to accomplish this. The problem with Command Prompt is that there does not seem to be a way of displaying files in "human" numerical order unless the files were first renamed with zero-padded filenames (which would be very time-consuming because I have hundreds of folders with hundreds of subfolders full of images). In other words, there is no "logical sorting" or "natural sorting" as it may sometimes be referred to as. Once you type "dir" in Windows Command Prompt, you'll get a result like this:

1.jpg
10.jpg
11.jpg
12.jpg
2.jpg
3.jpg
4.jpg
5.jpg
6.jpg
7.jpg
8.jpg
9.jpg

...whereas "natural sorting" would give you a result like this:

1.jpg
2.jpg
3.jpg
4.jpg
5.jpg
6.jpg
7.jpg
8.jpg
9.jpg
10.jpg
11.jpg
12.jpg

As you can notice, the problem is that 10-12 are displayed before 2-9 in Windows Command Prompt. Thus, if the files are renamed with a batch for Windows Command Prompt, the lack of any "natural sorting" will inevitably affect the renaming. In Bash, however, this problem can be bypassed by piping your preferred directory listing command through "sort -h". For example, "ls|sort -h" displays all the files in "human" numerical order.

The problem is that I'm a complete beginner with Bash, and I did my utmost to put together a shell script but everything was to no avail. I also ran many different ready-made scripts (with some minor changes) without any success. For example, I found the shell script below on Stack Overflow but it doesn't do quite what I want it to do even though I did my best to edit it:

find -iname '*.jpg' | sort -h # works only for .jpg files ATM...
gawk 'BEGIN{ a=1 }{ printf "mv %s %04d.jpg\n", $0, a++ }' | # build mv command
bash # run that command

Do you have any suggestions on how to remedy this problem with Bash so that I could rename my image files in the way explained above? I really wanted to do this by myself but I don't have the skills so any help you can provide would be most welcome!

P.S. The images are on my Windows PC but I'm using Cygwin (a Unix-like environment and command-line interface for Windows) to run all shell scripts for Bash.

This kind of problem always seems so easy until you try. I've made many attempts, but they're never quite satisfactory. I haven't found any kind of general solutions that humans can't break, and trying to fix up the corners just makes it easier to confuse.

Imagine this sequence of files for instance.

file1.jpg
file2.jpg
...
file8.jpg
file9.bmp
file10.jpg

Is it part of the sequence, or is it not? It's ambiguous.

Or how about this one:

cover.jpg
asd1.jpg
asd2.jpg
asd3.jpeg

Does cover belong at the beginning or the end, and how would you tell a program to know this? Are you sure the cover will always belong at the beginning?

Should we assume the jpeg is a typo because they have the same prefix of asd? Where does the prefix start, where does it end, and how is it deliniated from the rest?

Or this sequence:

cover for some file name containing 1234.jpg
some file name containing 1234 #1.jpg
some file name containing 1234 #1.jpg
some file name containing 1234 #1.jpg
...
some file name containing 1234 #365.jpg

Do 'cover for some file name containing 1234' and 'some file name containing 1234 #' belong to the same sequence or not? Should # be considered part of the name?

And what happens when you get a sequence like this?

1a.jpg
1b.jpg
1c.jpg
1.jpg
2.jpg
3.jpg
4.jpg

You end up tacking on special case after special case after special case onto your program to handle all these special things -- and then you get something like this thrown at you:

file1a.jpg
file2a.jpg
file3a.jpg
file4a.jpg

Or maybe the files all have 'cover' in their name so they all end up being 'special' and jammed in first in no particular order, or even skipped. Or this, or that, or a million other things.

If humans would just name their files in rational, consistent, predictable ways, they wouldn't need reordering. But all too often human context is needed to explain where things belong.