In one paragraph
Image buttons must have alternative text means a tested input of type image lacks an accepted accessible name. Describe the action the control performs, not merely the picture used to draw it. Prefer an appropriate non-empty alt for this native element and verify any other naming sources agree. An empty decorative alternative alone is not suitable for a meaningful submit control.
Why it matters
An image input is a form control as well as a graphic. A magnifying-glass illustration may mean Search, while a payment graphic may submit an order. Users need that action identified before activating it through keyboard, speech, or another input method.
Do not transfer the decorative-img pattern blindly to active image controls. A purely decorative ordinary image can use an empty alternative, but a meaningful action needs a name from a supported mechanism. The exact spoken output varies rather than always being one fixed button announcement.
An alt presence test does not judge whether the action is accurately described. A non-empty label such as Image or Click here can remain misleading. A payment name should not imply a harmless preview when it commits a transaction.
How WebAuditOS evaluates this finding
The supplied input-image-alt rule tests applicable rendered image inputs. It accepts non-empty alt and documented ARIA, title, or native-label naming paths. The ordinary value attribute is not the action-name path this rule relies on.
Critical impact maps to high severity. Hidden or unsampled instances can go untested, and a pass does not evaluate label accuracy, target size, focus, or the complete form process. Other native button forms have separate naming rules.
How to fix it
- Identify the actual submitted action and whether an image input is necessary. Check the form endpoint, button name, and relevant confirmation before writing the alternative.
- Add useful action wording. For example:
<input type="image" src="/images/search-button.png" alt="Search" name="searchSubmit">Use a real suitable asset and the application's intended behaviour. The example does not label the search field or implement results.
- Keep the alternative consistent with any text drawn into the graphic and the computed accessible name. Review explicit ARIA overrides that might replace an accurate alt with unrelated wording.
- If the image is only decoration around a normal action, consider a native button with visible text and appropriate decorative artwork instead. Do not change the element without understanding the submission contract.
- Preserve image-input coordinate semantics where the backend uses them. Replacing an image input with a different button can change submitted x/y fields or action routing even when the screenshot looks similar.
- Test asset failure, keyboard activation, focus visibility, validation, and responsive presentation. A transparent or broken graphic should not leave an otherwise meaningful action impossible to identify.
- Fix repeated templates and translated asset versions. A correctly named desktop image should not coexist with an unnamed mobile alternative or an untranslated action graphic.
- Rescan and manually verify the action name against what actually happens. Review neighbouring ordinary button, image-alt, and form-label results separately.
What the fix can break
Changing a name or submit value can affect tests and backend action routing. Treat functional data and human-readable wording deliberately rather than conflating them.
A graphical label can remain hard to read or scale even after alt is supplied. Consider visible real text and appropriate contrast and resizing design.
Do not add a presentational role to an active image input to escape naming requirements. Preserve its meaningful control identity.
Questions
Is empty alt always enough because the picture is decorative?
Not for an otherwise unnamed active image button. Its action still needs to be communicated.
Does value=Search provide the expected name here?
The supplied rule does not accept value as its image-input naming path. Use the appropriate native alternative and inspect computed output.
Can I replace the image input with a button safely?
Often, but verify submission data and behaviour, including coordinates or action identifiers used by the server.
Which related checks help?
Review button names, image alternatives, and the automation limits guide.
What severity does this get in the report?
High. An axe finding takes its severity from the impact axe reports for the element that failed, and the report maps critical and serious to high, moderate to medium, minor to low. This rule is declared critical in axe-core, so a failure arrives as high.