How does SiteSeeker index images?
11/5/2012SiteSeeker can in addition to HTML and other frequently occurring document types also index images.
Images are indexed by words in their URLs, by words in link texts (unusual) and also by alt and tile attributes in the img tags. Add more descriptive words in the alt text in order to get better search results on images, for example, like this:
<img src="car.jpg" alt="A picture on a yellow Volvo S60 with turbo and alloy wheels" title="A car">
For optimal results, formulate the alt text as a sentence since the website visitor may see the text while the image is loading, (the title attribute is displayed while hovering with the mouse over the image). Do the following if you want to add more keywords.
<img src="car.jpg" alt="A picture on a yellow Volvo S60 with turbo and alloy wheels (car, Volvo, yellow metallic, turbo, alloy wheels)" title="Volvo S60"
SiteSeeker will then take into account all words, without them being intrusive to the visitors since the keywords are within parentheses. Words in the adjacent headline and document title from the page where the image is located is associated with the images.
Image search can be improved on by allowing SiteSeeker to index meta information about the images.
Tip: it is important to choose relevant filenames, since words in the filenames of the images are indexed, (if the filename is part of the URL). For example, the file names CarVolvoTurbo.jpg or car_volvo_turbo.jpg are better than carvolvoturbo.jpg since SiteSeeker can identify and index the individual words in the first two examples but not in the last. The rescaled images are displayed in the relevant size in the results list. The user can then choose to view the image in the normal size or from the image's source document.
Filtering of unwanted images
SiteSeeker automatically filters low resolution images and images used as graphical elements on the website. You can adjust the minimum size for indexed images, the default setting for the lowest image size is 4 kB.
If an image is filtered from the search due to its size, you can manually include it in SiteSeeker Admin. Enter the URL of the image in the Server settings for the appropriate server, Indexing restrictions > Allow paths. Note that the URL may not end with an asterisk (*).
