[PDFBOX-5196] Wrong color space detected for some Jpeg images

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.23
• Fix Version/s: 2.0.24, 3.0.0 PDFBox
• Component/s: Parsing
• Assignee: Tilman Hausherr
• Reporter: Stéphane Vallée
• Created: 20/May/21 20:54
• Updated: 11/Jun/21 09:08
• Resolved: 22/May/21 05:47

Description

When retrieving the number of components in a jpeg image, class org.apache.pdfbox.pdmodel.graphics.image.JPEGFactory seems to retrieve the wrong metadata ("markerSequence/sos/@numScanComponents" instead of "markerSequence/sof/@numFrameComponents").

As a consequence, a wrong color space might be assigned to the image, preventing such image to be properly rendered in an PDF document.

Attached is a jpeg file whose color space will be wrongly determined (PDDeviceGray.INSTANCE instead of PDDeviceRGB.INSTANCE)

 

 

 

Issue Links

• duplicates: PDFBOX-5137 Wrong classification of an JPEG image leading to a blank image added to a pdf document