What Is Visual Regression Testing? Top 5 Visual Testing Tools

Brian
9 min readDec 8, 2022

What Is Visual Regression Testing?

Visual regression testing is a type of testing employed to confirm that recent program or code change in the software or application has not adversely impacted its Graphical User Interface (GUI). Visual regression testing ensures that all layout visual elements, (i.e. buttons, menus, icons, text, fields, etc.) are displayed as expected across all types of browser, device, and platform available. With well-executed visual regression testing, developers can be confident that all end users will have a perfect visual experience when using the application.

Visual regression testing is a form of regression testing. Regression testing ensures that the software or application does not fail after a code modification. Visual regression testing is the same, yet its focus is more on validating the aesthetic side of the software.

There are a lot of elements to consider when performing visual testing, including elements’ size, width, length, position, executability, visibility, readability, alignment, and consistency across all screen resolutions.

Why Is Visual Regression Testing Important?

Without visual regression testing, we run the risk of visual bugs going undetected, slipping into production and eventually into the application GUI.

The GUI is what an end-user sees and interacts with. If the application’s GUI is too buggy, the user will have a hard time navigating your product, leading to a poor User Experience.

Poor User Experience creates a negative ripple effect on the company’s bottom line. They will be driven away by the lack of aesthetics and usability of the product, leading to lost sales opportunities.

Poor GUI also equates to a lack of professionalism. Nothing screams “bad quality” more than a misplaced button or a terrible text layout. User Experience impacts how users view your product and even your brand. If you are trying to establish a market presence, it is crucial to maintain this sense of professionalism through proper visual regression testing.

visual validation testing common visual bugs
An Emergenc. Source: Reddit
visual regression testing UI bugs
Falsely displayed text. Source: Also Reddit
visual regression testing common visual bugs
Misplaced text and buttons on Google Map. Source: Yes, Reddit.

The images above demonstrate the catastrophe of no UI testing.

Skipping visual UI testing in your test plan is detrimental for 1 more reason: functional testing can not catch visual bugs.

Functional testing can only validate if an element works. A misplaced button surely works, but it is not in the right place, and functional testing can’t recognize that, unless the testers create uniquely coded assertions and selectors for every single element to identify their locations.

This practice is not scalable. As your software scales, the number of elements to code increases exponentially, and it is time-consuming to maintain them. 1 single change in the code requires revising the entire test code.

Visual regression testing tools employ advanced algorithms and technology to “scan” the GUI, imitating the way our human eyes see the elements on the screen, producing more accurate results for less effort.

Types Of Visual Regression Testing

Manual Visual Regression Testing

Manual visual regression testing is the traditional approach to visual testing, using no tools but solely the testers’ visual acuity to spot the bug.

However, we are fallible humans and we make human errors. That is not to mention the number of screen sizes, display resolution sizes, and browsers that need to be verified, which makes this approach obviously counter-intuitive.

visual testing tools using snapshot testing comparing baseline images
Can you spot 9 different spots in these 2 images? Try it, record your time, and imagine doing that repeatedly for hundreds of browsers and devices

That does not mean manual visual regression testing is unnecessary. When performing ad-hoc testing or exploratory testing, manual testing is faster and easier. For large-scale testing, we need to automate.

Automated Visual Regression Testing

Automated Visual Regression Testing gives the human eye a break by using automation frameworks to automate the process of finding and comparing visual elements across all browsers and devices to spot visual regressions. Common visual testing automation frameworks include Selenium WebDriver, Cypress, or Appium.

However, automation testing has its drawbacks. Since automated UI test scripts rely on using selectors, they have to be perpetually revised when the code evolves.

Pixel-by-pixel Comparison

A classic approach to automated UI testing is snapshot testing (or pixel-by-pixel comparison), which involves comparing a real-time snapshot against a baseline snapshot at a pixel level and flagging all changes to be revised. This approach sounds great, yet it does not take into account a few aspects:

  1. There are dynamic elements on a website, such as the Cart icon. It changes based on the actions the visitor takes. Automation scripts register this as a visual bug because the pixels are different.
Pixel-by-pixel Comparison — visual regression testing dynamic content on UI
Certain actions lead to changes in Cart icon on a webpage, but that does not mean it is a visual bug

2. Different hardware renders images differently, which can cause discrepancies in color and anti-aliasing, leading to false negatives

3. Padding/margin also varies across browsers and devices

AI can add another layer of intelligent comparisons to this comparison with Layout Comparison and other smart features.

AI-powered Visual Testing

AI Visual Testing can distinguish between critical and acceptable visual changes. For example, testers can configure the script to ignore dynamic elements and regions on the UI. They can also set up the acceptance level of discrepancy between snapshots, so that negligible change is not flagged as a bug.

A relatively new and advanced visual testing feature is Layout Comparison, which defines and matches various zones between the checkpoint and baseline snapshot and highlights any visible layout changes. Another feature is text/content-based comparison, which leverages Machine Learning to extract and compare text on the screen regardless of font size, family, and color and identify truly modified text from the original version.

The best part of Machine Learning is that it can gradually learn the way you structure your application from a content and layout perspective. As your code evolves, the machine gets better, generating ever-improving test results that are relevant to your business needs.

How To Choose A Visual Regression Testing Tool

The success of a visual testing project depends so much on the tool we choose. A good visual testing tool has to provide all of the features you specifically need, allows you to scale, and has a suitable and flexible pricing plan that fits your team size and test release frequency.

These are the questions that you should ask yourself when choosing your visual testing tool:

  • Can Your Team Utilize The Tool? Automation UI testing tools require significant coding knowledge, especially open-source tools. This technical barrier is challenging for QA teams with limited IT backgrounds. For these teams, it is advisable to choose low-code/no-code automation tools.
  • What Is Your Team Budget? Automation visual testing tools require an upfront investment. Consider your budget over a long timeframe before making the decision to choose either an open-source or commercial solution.
  • What Are Your Testing Needs? The aspects to consider vary across teams: What are the browsers/devices you need to test? Is your UI frequently updated with many dynamic content/elements? What is your release frequency?

Top Visual Regression Testing Tools To Catch Visual Bugs

To save your time doing research, I have compiled this list consisting of the 5 best visual regression testing tools on the market. They are all amazing tools with distinct features. With this list, you will have an easier time deciding which solution to use after considering all of the aspects of your testing needs.

Paid Tools

  1. Applitools
Applitools — visual testing tools

Applitools is the most popular Visual Testing tool on the market. Having entered the market relatively early, it changed the visual testing landscape by introducing the scanning technology to humanize the testing scripts and shorten the product release cycle.

Highlighted features:

  • AI-powered computer vision algorithms that emulate the human eye, analyzing and reporting only the most visible differences
  • Scalable across a wide range of devices and browsers
  • Support multiple testing frameworks
  • Integration with leading CI systems and collaboration services

Website: https://applitools.com

Price: Applitools offers Forever Free account for 1 User & 100 Checkpoints per month to perform functional and visual testing. For advanced features, they have 3 Tiers: Starter, Eyes, and Ultrafast Test Cloud. You can contact their Sales team for a detailed quote.

2. The Katalon Platform

Katalon — visual testing tools. All-in-one testing platform

Katalon Platform prides itself as the All-In-One quality management platform built on Appium and Selenium, allowing you to perform functional testing and visual testing in 1 single place, saving your team tremendous time. Katalon Platform has many visual testing features to help your team achieve visual quality along with usability.

Highlighted Features:

  • Offers centralized platform to perform both functional and visual testing
  • AI-powered features (Layout Comparison, Text-to-text Comparison) to reduce false positives
  • Comprehensive documents, tutorials, and forum support

Website: https://katalon.com

Price: Katalon Platform offers Forever Free usage. For more advanced features and quota, Katalon offers an attractive annual pricing plan.

3. Percy (from BrowserStack)

Percy (BrowserStack) — visual testing tools

Percy is a visual regression testing solution from BrowserStack. Percy allows you to perform intelligent visual testing across browsers and devices. It can make smart comparisons, groups those visual changes together, then selects only the most noticeable visual regressions. They also have a comment and notification feature to help your team stay updated on the project progress.

Highlighted Features:

  • Automatic status updates in pull requests whenever visual changes are detected
  • Snapshot stabilization technology automatically freezes animation to minimize false positives
  • Ignore specific areas on the page
  • Seamlessly integrate into existing CI/CD workflow
  • Support popular test automation frameworks and languages

Website: https://www.browserstack.com/percy

Price: Percy offers a Free package with unlimited users at 5000 screenshots per month. For more advanced features and quota, Percy offers a Desktop as well as a Desktop & Mobile visual testing package at $149 and $199 per month, billed annually.

Open-source Tools

4. Selenium

Selenium — visual testing tools

Needless to say, Selenium is the leader in today’s open-source testing tool market. Being compatible with a wide range of programming languages, testing frameworks, browsers and operating systems. They bring testers and developers unparalleled testing flexibility that many other test automation tools can’t easily provide. However, Selenium does not have built-in image comparison that is vital to visual regression testing, so a third-party library must be used.

Highlighted Features:

  • Unmatched testing capabilities for teams of all sizes
  • Support a wide range of languages and frameworks on many platforms with large learning resources from the global community
  • Seamless CI/CD integration
  • Parallel test execution to significantly reduce your testing time and control

Website: https://www.selenium.dev/

Price: Open-source

5. SikuliX

SikuliX — visual testing tools

SikuliX is an open source tool that uses Image Recognition and Control GUI technology to identify and interact with GUI components, such as identifying a certain button on the screen, then controlling the mouse and keyboard to interact with the identified button. This makes SikuliX extremely handy when you don’t have access to the source code of the application or web page you want to test.

Highlighted Features:

  • Can be used to automate Flash Objects/Flash Websites
  • Can perform visual testing for web pages, desktop applications on Windows, Linux, Mac, iPhone, and Android
  • Can automate certain tasks on the GUI even when you don’t know the ID of the elements
  • Can easily be integrated into other third-party applications

Website: http://sikulix.com/

Price: Open-source

— — — -

Visual regression testing is a great practice to bring only the best and perfect GUI to your users. When manual testing and automated UI testing with open source tools become unscalable, it is necessary to invest in a low-code solution that goes along with your product in the long run.

The list of tools I introduced certainly was not exhaustive, but it still represents some of the market leaders in the visual regression testing landscape. They all have AI/ML to overcome the drawbacks of writing visual testing scripts and expand the capabilities of our human eyes. Organizations can now deploy confidently with the help of these quality solutions.

--

--

Brian

Testing evangelist. Automation advocate. Selenium committer. Experienced with open source, UFT, Katalon, and others.