The use of surface acoustic waves (SAW) in cell biology has gained high attention in the past years. Previous works show that SAW treatment of artificial wounds in vitro can accelerate wound healing by up to +135%. However, little is known about the mechanobiology behind these effects, and a stimulation has only been proven for continuous SAW signals so far. We here show that the stimulation efficacy observed in previous studies is preserved for pulsed stimuli applied to Madin-Darby canine kidney (MDCK-II) cells in wound healing assays on SAW chips at a resonance frequency f SAW = 160 MHz. Moreover, for a reproducible and reliable image analysis, we present the SegFormer-based deep learning algorithm Neural Cell Edge Detector for the cell edge segmentation and image binarization, that allows an automated determination of the stimulation efficacy. With these tools, we explore a wider range of applicable SAW intensities up to P in = 21 dBm (128 mW) with a maximum stimulation efficacy of E = 201%. We show that the order of magnitude of the stimulation effect is reproducible under the variation of SAW signal characteristics power P in, duty cycle D and pulse width τ, while the mean energy over time is constant. Below a distinct pulse duration limit τ = 100 ms, no direct stimulation effect was observed. From a mechanobiological point of view, these findings and time scales could narrow down the potentially triggered cellular mechanisms during the stimulation.