I appreciate that you want to get to the bottom of this, but I think the energy cost of restarting the engine is so small that a scangauge could not accurately measure it.
Look at it this way, I can bump start the engine in my 2000 lb car at 5mph and wind up at, say, 4mph. I'm quite certain I could get back up to 5mph with the fuel that would have been spent idling for 10 seconds.
For those of you who intend to measure injector pulsewidth, you will also need to measure how frequently the injectors fire, and you should measure each injector to determine if they're all firing at once or individually.
I don't know about other cars, but I know my car very well. When the engine is started all 4 injectors fire together, it's called 'batchfire' and adds extra fuel for a very brief time, even when fully warmed up. All 4 injectors fire every 1/2 revolution. Once the engine is started the ecu switches to firing the injectors individually, once per 2 revolutions. I imagine other port fuel injection engines may be managed like this.