Stream processing to solve image search by similarity
The classic use of Stream Processing platforms enables working with data in real time, which allows you to generate data analysis quickly attending to a decisionmaking process. However, you can use these platforms for other applications such as indexing and subsequent use of similarity search object...
| Autores: | , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2015 |
| País: | Argentina |
| Institución: | Universidad Nacional de La Plata |
| Repositorio: | SEDICI (UNLP) |
| Idioma: | inglés |
| OAI Identifier: | oai:sedici.unlp.edu.ar:10915/50180 |
| Acceso en línea: | http://sedici.unlp.edu.ar/handle/10915/50180 |
| Access Level: | acceso abierto |
| Palabra clave: | Ciencias Informáticas stream processing metrics spaces MPEG-7 sparse spatial selection Metrics |
| Sumario: | The classic use of Stream Processing platforms enables working with data in real time, which allows you to generate data analysis quickly attending to a decisionmaking process. However, you can use these platforms for other applications such as indexing and subsequent use of similarity search objects in a database. The images can be displayed on a metric space, which has features that allow rules to discard a not similar image quickly without making costly computations. This paper presents the use of a Stream Processing platform to index images generated by different users. For this, it is necessary to represent these images by vectors containing different MPGE-7 features. This paper shows a Stream Processing platform using its processing elements (PEs) in parallel to speed up the operations involved in the index construction. |
|---|