Dicoogle, a Pacs Featuring Profiled Content Based Image Retrieval

Abstract

Content-based image retrieval (CBIR) has been heralded as a mechanism to cope with the increasingly larger volumes of information present in medical imaging repositories. However, generic, extensible CBIR frameworks that work natively with Picture Archive and Communication Systems (PACS) are scarce. In this article we propose a methodology for parametric CBIR based on similarity profiles. The architecture and implementation of a profiled CBIR system, based on query by example, atop Dicoogle, an open-source, full-fletched PACS is also presented and discussed. In this solution, CBIR profiles allow the specification of both a distance function to be applied and the feature set that must be present for that function to operate. The presented framework provides the basis for a CBIR expansion mechanism and the solution developed integrates with DICOM based PACS networks where it provides CBIR functionality in a seamless manner.

A Multimodal Search Engine for Medical Imaging Studies

“…Dicoogle's engine has since been augmented with CBIR [8], supporting automatic image feature extraction on indexation, as well as similarity metrics for performing query-byexample. The concept of CBIR profile was introduced in order to cope with the rapid appearance of new feature extraction and similarity techniques that may only be compatible with the content of a certain modality.…”

“…This paper presents a platform for making multimodal searches in a medical imaging repository, supporting complex queries composed by the combination of textual and visual information. The proof of concept was developed using Dicoogle, an open-source PACS archive, and an existing CBIR platform [8], thus withholding existing contributions in a decoupled and modular fashion. The result is a highly flexible architecture for executing multimodal searches in a PACS, for benchmarking and for clinical use, as well as a web-based platform that addresses functionality and usability concerns.…”

Anatomy of an Extensible Open Source PACS

“…With the project growing organically, a number of forks and branches were created providing distinct functionality. Dealing with multiple implementations [28][29][30][31][32] quickly became unattainable and cumbersome to scale. The merge of branched code with the stable branch has proven not to be free of issues and led to minor and sometimes major duplication of code having slightly different interfaces, to the need to adapt existing functionality to changes introduced elsewhere in the code, or even caused the introduction of race conditions.…”

“…Given the very active research effort in PACS systems, we felt there is a need for an environment where an idea may quickly be prototyped, tested, and validated. On the other hand, the addition of a CBIR module supported by a separated indexing engine [30] and plans to leverage cloud-based data storage required us to re-evaluate our previous assumptions (such as the internal interfaces to query and indexing based on Lucene's API). This led us to refactor the application's architecture, with the goal of streamlining and easing third-party development, a process that benefited from the field expertise obtained from previous iterations and deployments of Dicoogle.…”

“…We will be analyzing the implementation of this mechanism as an extension for Dicoogle using the plugin system mentioned before. A detailed analysis of the strategies and algorithms employed can be found in [30]. The objective of Dicoogle CBIR plugin (Fig.…”