It looks like the vertical axis is okay, but the horizontal one is off. If those images are of two different attempts, then it probably isn't lost steps, since the results seem consistent while lost steps are random. I'd look at the "steps per unit" setting on your horizontal axis. You get that by multiplying the screw pitch by the steps per revolution in the motor (usually 200) by the microstepping factor of your drive. If you need more help with this, further details about your setup would be useful.