We provide the candidate image value in the 'CandidateImageData' tag for word documents only, not for the PDF files. So, you will not be able to get the candidate image data value for a PDF file.
Also, in the document, we do face detection to get the candidate images from all of the images.
Suppose the candidate has mentioned multiple images. then the image with the face will be extracted as a candidate image. Other than the candidate, face images will not be extracted by the parser. In short, the face is compulsory and should be there in the image.
Also please check your API Settings and make sure that HTML Format is enabled.
If you are parsing the doc/docx file and still not getting candidate image data, you can contact RChilli Support via creating a ticket at RChilli HelpDesk or sending an email at email@example.com. The Support Team will assist you further.