Surgical Planning Laboratory - Brigham & Women's Hospital - Boston, Massachusetts USA - a teaching affiliate of Harvard Medical School

Surgical Planning Laboratory

Image Archive

2013 :: 2011 :: 2010 :: 2009 :: 2008 :: 2007



Lung Extraction, Lobe Segmentation and Hierarchical Region Assessment for Quantitative Analysis on High Resolution Computed Tomography Images. Region and Type Segmentation. Top row shows example results from the automatic processing stage. Bottom row shows lobe segmentation results produced with the interactive lobe segmentation tool. Read more here


Local White Matter Geometry Indices from Diffusion Tensor Gradients. Streamline DT tractography colored by dispersion index (a,c) and curving index (b,d). Red indicates high index values. The T2-weighted image is shown in the background for reference. Note the consistent behaviour of the pre-computed dispersion and curving indices with respect to the tractography results. The regions in the green boxes are discussed further in the text. Some of the strong dispersion regions in the internal capsule (c) are due to fibre fanning orthogonal to the image plane. Read more here


Computational Neuroanatomy: Ontology-based Representation of Neural Components and Connectivity. Image atlas of the brain. Image atlases represent spatial information by providing a parcellation of the anatomic structures contained in the brain (left). Each structure is represented as a spatial region of uniform color. Other anatomic knowledge about the structure, such as functional information, is not represented. Image atlases are generally used to infer the anatomic localization of brain structures in individual subjects by registering their images to the atlas. For example, the anatomic identity of areas of activity in fMRI are identified in this manner (right). Read more here


A role for self-gravity at multiple length scales in the process of star formation. CLUMPFIND featureidentification algorithms as applied to 13CO emission from the L1448 region of Perseus. a: 3D visualization of the surfaces. b: CLUMPFIND segmentation. A very large number of clumps appears, because of the sensitivity of CLUMPFIND to noise and small-scale structure in the data. Read more here