Visual Verifications

Visual Verifications Overview

There are two ways to perform Visual Verifications using Functionize's Computer Vision feature:

We employ a machine learning approach to compare images, making the process faster and taking into consideration trivial changes that can occur within the browser. Just imagine the browser rendering an image off by a few pixels in each execution. Our baselines are established per browser to ensure that browser-specific false negatives do not occur.

 

Both of these visual verifications operate similarly:

  • Match Requirement Default % set to 92%
    Your project has a default match requirement % for both types of verification. We keep this tolerance in place for common variations, but, of course, you can adjust this value higher or lower. The default value will be passed into Architect and can be overwritten on an action-by-action basis when creating the test or from the UI after the test has been executed.
  • Baselines
    Baselines for the image are established during the first successful run, per browser. However, Element baselines are set to Chrome's default when creating a test with Architect.

Visual Verification: Full Page Verification

A full-page visual verification, referred to as a Visual Page Check, allows you to verify that an entire screenshot, even when not in the current viewport, always matches your baseline. There are two types of Full-Page verifications:

  • Baseline Verifications: In every execution, we compare against the baseline, which is set for Chrome via Architect during creation or during the first successful execution of the test. The baselines may be updated after any execution to match the current screenshot.

  • Previous Step Verification: This should be used when you want to verify a screenshot against a previous step's verification. For example, if you would like to compare a screenshot with the previous screenshot, you might use the Previous Step Verification method.

Visual Verification: Element-Based Verification

Element-based verifications are a feature of any Verify Action. The element could be an image, but also a div, span, or any kind of element that you would typically baseline. On the Slider View screenshot, the area highlighted in green for the selected element will be used as the image for comparison. There are two types of Element-based verifications:

  • Baseline Verifications: In every execution, we compare against the baseline, which is set for Chrome via Architect during creation or during the first successful execution of the test. The baselines may be updated after any execution to match the current run's element.

  • Previous Step Verification: This should be used when you want to verify an element against a previous step's verification. For example, if you would like to compare a company logo that always appears just like the first step's logo, you might use the Previous Step Verification method.