1. How easy is it for me to evaluate Imprezzeo?Very easy! The evaluation version - Imprezzeo Standalone - is supplied as a self extracting zip file which contains H2 database as the repository; it also ships with a simple user interface, all you need to do is provide images. It is recommended though to limit the size of an evaluation set to 250,000 images. Click here to complete an evaluation request form.
2. How many images do I need as a minimum for Imprezzeo to work?A minimum of a few hundred images is all that is required for Imprezzeo; however, it is optimised for larger image libraries so we recommend a minimum of 20,000. The systems are extensible and scalable, and we have worked on libraries containing over 20 million images.
3. How do I see the search results?Imprezzeo ships with a RESTful web interface, allowing results to be displayed via the user’s browser.
Reports can also be provided as downloadable PDF files.
4. Do I need to tune Imprezzeo to optimise my search results?No, Imprezzeo comes with two pre-tuned configurations that work well for many collections:
However, collections vary and Imprezzeo has a toolkit that allows fine-tuning of results.
- Creative - for artistic or creative collections
- Editorial - for news and media-type collections
5. How do I get the best performance out of Imprezzeo?You can uniquely adjust the image comparison settings within Imprezzeo. Imprezzeo allows you to specify different weightings against the attributes of an image. This allows optimisation of various search parameters, such as weightings for edge, scalable colour layout, texture, etc. and optimising aspects of face recognition.
6. How well does Imprezzeo image search work when scaled up?Imprezzeo is designed to perform at scale and can manage libraries of tens of millions of images. It does this by ingesting images and creating an Imprezzeo “fingerprint” file (approx. 1Kb each) for every image.
Imprezzeo’s architecture is designed to support multiple tiers, from small images collections (<500,000 images), to large collections (<20 million images) to very large collections (>20 million images). Imprezzeo’s architecture allows our customers to scale seamlessly from tier to tier.
For large image collections, these “fingerprints’ are loaded into memory at start-up. This maintains performance by avoiding the delay caused by database and disk reading access each time a query is carried out. New images that are added to the Imprezzeo repository are available to the user immediately (and vice-versa, if an image is deleted from the repository, it is no longer available).
7. What file type and image resolution does Imprezzeo require?Imprezzeo can work with all images named using the UNC convention, e.g. jpegs, bitmaps, tiff and gif files. Images must be of
a minimum resolution to ensure quality comparisons can be carried out. Thumbnail images of a minimum of 320 pixels at its
longer length are usually sufficient.
8. What software platform(s) does Imprezzeo run on?The core Imprezzeo search engine runs on any platform that supports Java 1.6 or higher. Additional Imprezzeo features also require Java SWT toolkit. This includes Microsoft Windows 2000, XP, Vista, 7, Linux and Apple Macs. Imprezzeo requires a JDBC database driver and supports the following databases: PostgreSQL, MySQL and Microsoft SQL Server, Apache Derby, DB2 and Oracle.
Please see the following document for further software specification details:
9. What is the minimum system specification needed to run Imprezzeo?Please see Technical Resources section opposite and download Imprezzeo Architecture and Technical Specifications for minimum hardware specification.
10. Can Imprezzeo be networked?Yes, dependent on your platform. Drives that are locally mapped (Windows) or mounted (UNIX) should also work.
SMB and CIFS mounted shared drives can also be managed.