One of the most important things you can do before working with your application code is to clarify what capabilities within eVe can help you meet your applications business requirements. Determining the visual search functions you want to employ will speed up eVe integration and application deployment while ensuring that you address your customers needs.
By determining the requirements of your application, you can ensure the best use of eVe to address your customers' functional requirements. During use case modeling and requirements review, you can nail down users' behavior in great detail; software behavior is dictated by user requirements.
With eVe, some of the main visual search capabilities you can enable within your application include:
|
eVe Feature |
High-Level API |
|
multiple search types |
Use eVe to make your application capable of performing a visual search based on any or all of the following properties of an image: texture, shape, color, or object region. You can allow users to search on one property at a time, or perform a search based on a combination of properties (such as color + shape). Additionally, you can enable users to specify the relative importance of each of the four basic properties within a search.
Determining how a user of your application will initiate and define a search is critical to successfully integrating eVe. |
|
partial image search |
Use this feature to enable users to select one or more regions within an image to use as search criterion. Because of the advanced segmentation technique built into eVe, partial image searching enables users to identify specific objects within an image and accurately search for both a whole image and for parts (or objects) of an image.
If you choose to integrate this feature within your application, see the Code Samples chapter for an example of how to incorporate it into an application. |
|
visual vocabulary |
A Visual Vocabulary is a set of images that is representative of a larger set or collection of images. For example, if you have 1000 images of zebras, you can program eVe within your application to select a subset of zebra images from those 1000 images that are most representative of the entire collection.
If you choose to integrate this feature within your application, you can enable users to base their searches on an image from a Visual Vocabulary rather than on a text-based entry. eVe allows you to programmatically automate the generation of a Visual Vocabulary for a database of images, and/or allow users to manually create their own vocabularies. |
|
similarity ranking |
eVe provides the SearchResults interface for creating, manipulating, and updating the results of a visual search. Within your application you can use eVe to sort, combine, truncate, rank, and append the results of a visual search.
How you want to manipulate the results of a visual search will determine what methods from the SearchResults interface you will implement. |
|
multiple platform support |
eVe consists of 100% JAVA APIs. This allows you to quickly and easily integrate eVe into your new and existing applications on many different OS platforms.
Check out the Installation Guide for a list of the Operating Systems that eVe supports. |
|
automatic asset ingestion |
You can use eVes API to set up automatic generation of visual metadata, proxy, and thumbnail images for your media assets. This allows you to ingest, analyze, and index assets automatically. |
|
database independence |
eVe supports flat files, ObjectStore, and Oracle. You can choose the optimal data storage environment for your application. |