ImageJ uses Java so will not run on an iPad. You should be able to use a Surface but you will need to take it out of tablet mode to run the program.

Any touchscreen laptop that can run ImageJ would work too, but a tablet is a little easier to handle.

What do you mean exactly when you say :

"I feel it is an inefficient and time-consuming exercise to analyze images using ImageJ on computers; "

Just a stupid question but you know ImageJ can be scripted right? Have you tried to automate some of the time-consuming operations you are doing manually by scripting them?

I agree with Bruno in that the use of scripting/automating saves a lot of time. For example, I process a lot of histology images, and I separate stains using the color deconvolution tool (https://imagej.net/Colour_Deconvolution) and then threshold. And using the macro recorder makes automation easy, and I can batch process images.

However, I have found that scripting doesn't work so well in situations where the software cannot distinguish between structures. For example, having it measure the perimeter of a specific organ/structure in a section through the whole body is difficult. The human eye can differentiate the organ, but as it has similar staining properties to the tissues around it and is slightly different in shape in different sections, the computer makes bad decisions. To date, I haven't come across a good piece of script that can handle these situations (if you know of one, I would love to try it out!). Often you must hand trace around the structure of interest, which is indeed inefficient and time-consuming. Tracing with a mouse is a real pain, but a stylus on a tablet is much faster and easier. Once it is saved as an ROI though, the script can come in.