Screenshot comparison (Visual regression) - assertion step

We plan to add a separate step for screenshot comparison. The first execution will save a screenshot as a baseline, and then every next execution will be compared to this baseline screenshot. If the new screenshot is different from the baseline, the test will fail.

Possibly, this step should also have an option to

• exclude some random elements from the DOM before taking a screenshot

• option to decide how this comparison should be done (pixel perfect or percentage)

• option to verify screenshot only for specific element on the page (given selector)

Any other ideas around this feature are more than welcome.