Abstract-We propose a shape-based approach to curve evolution for the segmentation of medical images containing known object types. In particular, motivated by the work of Leventon, Grimson, and Faugeras [15], we derive a parametric model for an implicit representation of the segmenting curve by applying principal component analysis to a collection of signed distance representations of the training data. The parameters of this representation are then manipulated to minimize an objective function for segmentation. The resulting algorithm is able to handle multidimensional data, can deal with topological changes of the curve, is robust to noise and initial contour placements, and is computationally efficient. At the same time, it avoids the need for point correspondences during the training phase of the algorithm. We demonstrate this technique by applying it to two medical applications; two-dimensional segmentation of cardiac magnetic resonance imaging (MRI) and three-dimensional segmentation of prostate MRI.Index Terms-Active contours, binary image alignment, cardiac MRI segmentation, curve evolution, deformable model, distance transforms, eigenshapes, implicit shape representation, medical image segmentation, parametric shape model, principal component analysis, prostate segmentation, shape prior, statistical shape model.

W e propose a model-based curve evolution technique for segmentation of images containing known object types. I n particular, motivated by the work of Leventon, Grimson, and Faugeras [4], we derive a parametric model for a n implicit representation of the segmenting curve by applying principal component analgsis to a collection of signed distance representations of the training data. The parameters of this representution are then calculated to minimize a n objective function for segmentation. W e found the resulting algorithm to be computationally eficient, able to handle multidimensional data, robust to noise and initial contour placements, while at the same time, avoiding the need f o r point correspondences during the training phasc: of the algorithm. W e demonstrate this technique by applying it to two medical applications.

The 2009 Influenza A (H1N1) pandemic disproportionately affected the developing world and high-lighted the key inadequacies of traditional diagnostic methods that make them unsuitable for use in resource-limited settings, from expensive equipment and infrastructure requirements to unacceptably long turnaround times. While rapid immunoassay diagnostic tests were much less costly and more context-appropriate, they suffered from drastically low sensitivities and high false negative rates. An accurate, sensitive, and specific molecular diagnostic that is also rapid, low-cost, and independent of laboratory infrastructure is needed for effective point-of-care detection and epidemiological control in these developing regions. We developed a paper-based assay that allows for the extraction and purification of RNA directly from human clinical nasopharyngeal specimens through a poly(ether sulfone) paper matrix, H1N1-specific in situ isothermal amplification directly within the same paper matrix, and immediate visual detection on lateral flow strips. The complete sample-to-answer assay can be performed at the point-of-care in just 45 min, without the need for expensive equipment or laboratory infrastructure, and it has a clinically relevant viral load detection limit of 106 copies/mL, offering a 10-fold improvement over current rapid immunoassays.

Bacteria are an enormous and largely untapped reservoir of biosensing proteins. We describe an approach to identify and isolate bacterial allosteric transcription factors (aTFs) that recognize a target analyte and to develop these TFs into biosensor devices. Our approach utilizes a combination of genomic screens and functional assays to identify and isolate biosensing TFs, and a quantum-dot Förster Resonance Energy Transfer (FRET) strategy for transducing analyte recognition into real-time quantitative measurements. We use this approach to identify a progesterone-sensing bacterial aTF and to develop this TF into an optical sensor for progesterone. The sensor detects progesterone in artificial urine with sufficient sensitivity and specificity for clinical use, while being compatible with an inexpensive and portable electronic reader for point-of-care applications. Our results provide proof-of-concept for a paradigm of microbially-derived biosensors adaptable to inexpensive, real-time sensor devices.