Thank you, Vivka! Your code leaves everything unchanged. But what I need is not to highlight all instances of 'width=' and 'height=' followed by three-digit numbers.
As I suggested in the original post, I have a text file with several sequences of this type:

<p class=NonNrPar align=center style='text-align:center;line-height:
normal'><img width=453 height=255 id="Immagine 1"
src="E_DIME_ORIG_file/image001.jpg"></p>

I'm trying to write a code that will
1) Find part of the above string: from <img to ><;
2) create a variable that contains the string width=### in the Found string;
3) create a variable that contains the string height=### in the Found string;
4) Delete the Found string: from <img to ><.

Hope someone can help. Thanks!