Visual Regression — All Articles & Tutorials

Visual regression refers to any rendering deviation introduced unintentionally between two versions of an interface: a padding that jumps, a color that drifts, a component that collapses on mobile after a framework update. These regressions almost always escape unit and functional tests, since the DOM may stay strictly identical while the visible rendering degrades. Detecting these deviations requires a stable baseline screenshot, a deterministic capture and a diff capable of distinguishing real bugs from harmless cosmetic variations (anti-aliasing, animations, dynamic data).

This page gathers articles dedicated to the baseline-capture-comparison-validation cycle: how to build a reliable baseline, how to handle false positives caused by fonts or mobile pixels, how to integrate a manual diff validation workflow within a QA team. You will also find feedback on classic pitfalls (mass CSS rewrites, Angular or React migration, image CDN change) that turn a routine deployment into a visual bug hunt. Delta-QA is part of this discipline with a desktop and local approach, but the topic far exceeds the tool: it is above all a methodology that gets refined project after project, and these articles aim to share what really works in the field, regardless of the stack used.