You can use the following example to extract all the images (not annotations) contained in a given PDF document.

Document inputFile = new Document();
if (inputFile.Open(path, "") == 0) {
    for (int pageIndex = 0; pageIndex < 3; pageIndex++) {
        Page page = inputFile.GetPage(pageIndex);
        if (page != null) {
            page.ObjsStart();
            List<Integer> imageIndices = getImagesIndex(page);
            if(imageIndices == null) continue;
            for(int i = 0; i < imageIndices.size(); i++) {
                int randomIndex = imageIndices.get(i);
                if (page.ObjsGetCharUnicode(randomIndex) == -1) { // image
                    String format = "png";
 
                    PDFImage extractedImage = getImage(page, randomIndex);
                    if(extractedImage != null) { //extract the image to the file system 
                        ImageIO.write(extractedImage.getImage(), format, new File(imageExtractPath + 
                            file.getName().substring(0, file.getName().lastIndexOf(".")) + "_" + pageIndex + "." + format));
                    } 
                }
            }
            page.Close();
        }
    }
}
inputFile.Close();

/**
 * Iterates on the given page characters and return the index of the found images.
 * 
 * @param page the page to check, ObjsStart must be invoked first.
 * @return a list containing the indices of the found images.
 */
 public static List<Integer> getImagesIndex(Page page) {
     List<Integer> indices = null;
     int charCount = page.ObjsGetCharCount();
     for (int i = 0; i < charCount; i++) {
         if (page.ObjsGetCharUnicode(i) == -1) { // image
             if(indices == null) indices = new ArrayList<Integer>();
             indices.add(i);
         }
     }
     return indices;
 }

/**
 * get an image from the given page.
 * 
 * @param page the pdf page object, ObjsStart must be invoked first.
 * @param imageIndex the index of the image char in the given page.
 * @return PDFImage object containing the BufferedImage extracted from the page, and the image coordinates.
 */
 public static PDFImage getImage(Page page, int imageIndex) {
     PDFImage pdfImage = new PDFImage();
     pdfImage.setImageIndex(imageIndex);
 
     BufferedImage extractedImage = page.ObjsGetImage(imageIndex);
     //page is extracted upside-down, so flip it
     if(extractedImage != null) 
         extractedImage = flipImageVertically(extractedImage); 
     pdfImage.setImage(extractedImage);
 
     //get image coordinates
     float[] imageRect = new float[4];
     page.ObjsGetCharRect(imageIndex, imageRect);
     pdfImage.setImageRect(imageRect);
 
     return pdfImage;
 }

public static BufferedImage flipImageVertically(BufferedImage input) {
     BufferedImage flipped = new BufferedImage(input.getWidth(), input.getHeight(), input.getType());
     AffineTransform tran = AffineTransform.getTranslateInstance(0, input.getHeight());
     AffineTransform flip = AffineTransform.getScaleInstance(1d, -1d);
     tran.concatenate(flip);
     Graphics2D g = flipped.createGraphics();
     g.setTransform(tran);
     g.drawImage(input, 0, 0, null);
     g.dispose();
     return flipped;
 }

public class PDFImage {

private int mImageIndex;
private float[] mImageRect;
private BufferedImage mImage;
  
public int getImageIndex() {
    return mImageIndex;
}

public void setImageIndex(int mImageIndex) {
    this.mImageIndex = mImageIndex;
}

public float[] getImageRect() {
    return mImageRect;
}
 
public void setImageRect(float[] mImageRect) {
    this.mImageRect = mImageRect;
}
 
public BufferedImage getImage() {
    return mImage;
}
 
public void setImage(BufferedImage mImage) {
    this.mImage = mImage;
}
}