Application and data fusion of different sensor modalities in tomographic imaging

Abstract

A crucial aspect in tomographic imaging is the choice of the sensing modality. It depends on the intended application since different sensing principles are sensitive to different physical parameters. This paper gives an overview of sensing modalities that are frequently used in industrial and medical applications. Common issues of tomographic measurement techniques are a small amount of available data and inherently ill-posed inverse problems, resulting in limited achievable reconstruction accuracy. Multimodality tomography has been proposed to partly overcome these problems by exploiting the complementarity of the sensing modalities for improved imaging performance. Data fusion approaches proposed for tomographic imaging are reviewed in this paper. The potential of multimodal tomography is illustrated by means of an application example from medical imaging.Ü berblick verschiedener Ansä tze der Sensorfusion fü r tomografische bildgebende Systeme.Ein wichtiger Aspekt bei der Anwendung tomografischer Messmethoden ist die Wahl des Sensorprinzips. Dieses hä ngt stark von der Zielanwendung ab, weil verschiedene Sensorprinzipien auf unterschiedliche physikalische Parameter empfindlich sind. Diese Arbeit gibt einen Ü berblick ü ber hä ufig verwendete Sensorprinzipien fü r industrielle und medizinische Tomografie. Ü bliche Probleme tomografischer Methoden sind geringe verfü gbare Datenmengen und inhä rent schlechtgestellte inverse Probleme, was sich in begrenzter erreichbarer Rekonstruktionsgenauigkeit auswirkt. Die Fusion verschiedener physikalischer Sensormodalitä ten ist ein Versuch, ihre komplementä ren Eigenschaften zur Erreichung einer Rekonstruktionsqualitä t auszunutzen, die mit einem einzelnen Messprinzip nicht erreichbar wä re. Diese Arbeit gibt einen Ü berblick verschiedener Ansä tze der Sensorfusion fü r tomografische bildgebende Systeme. Ein Anwendungsbeispiel illustriert die mö gliche Verbesserung der Bildqualitä t.

Model-based Data Fusion in Industrial Process Instrumentation

“…Another option for the combination of two modalities is parallel processing of the totality of raw data. However, this raises serious issues of data association (Steiner, 2007). For the particular combination of sequential URT-ECT fusion the URT reconstruction can be used to deduce an outer approximation of the inclusion region containing the disperse phase of the material distribution, i.e.…”

A bio-electromechanical imaging technique with combined electrical impedance and ultrasound tomography

“…Potential alternatives that have recently received considerable attention are tomographic methods based on non-ionizing sensing fields that can be built from inexpensive components and could therefore be routinely used. Overviews of different available sensing modalities can be found in Natterer (2007) and Steiner (2007). These methods include near-infrared (NIR) optical tomography (Arridge 1999, Brooksby et al 2003, Zhu et al 2003 and electrical impedance tomography (EIT) (Cheney et al 1999, Cherepenin et al 2002, Choi et al 2007, Holder 2005.…”