Abstract: A method of fabricating a device includes forming a dummy gate over a plurality of fins. Thereafter, a first portion of the dummy gate is removed to form a first trench that exposes a first hybrid fin and a first part of a second hybrid fin. The method further includes filling the first trench with a dielectric material disposed over the first hybrid fin and over the first part of the second hybrid fin. Thereafter, a second portion of the dummy gate is removed to form a second trench and the second trench is filled with a metal layer. The method further includes etching-back the metal layer, where a first plane defined by a first top surface of the metal layer is disposed beneath a second plane defined by a second top surface of a second part of the second hybrid fin after the etching-back the metal layer.

Abstract: A method of forming a semiconductor device includes forming semiconductor strips protruding above a substrate and isolation regions between the semiconductor strips; forming hybrid fins on the isolation regions, the hybrid fins comprising dielectric fins and dielectric structures over the dielectric fins; forming a dummy gate structure over the semiconductor strip; forming source/drain regions over the semiconductor strips and on opposing sides of the dummy gate structure; forming nanowires under the dummy gate structure, where the nanowires are over and aligned with respective semiconductor strips, and the source/drain regions are at opposing ends of the nanowires, where the hybrid fins extend further from the substrate than the nanowires; after forming the nanowires, reducing widths of center portions of the hybrid fins while keeping widths of end portions of the hybrid fins unchanged, and forming an electrically conductive material around the nanowires.

Abstract: A semiconductor device includes a first dielectric layer, a stack of semiconductor layers disposed over the first dielectric layer, a gate structure wrapping around each of the semiconductor layers and extending lengthwise along a direction, and a dielectric fin structure and an isolation structure disposed on opposite sides of the stack of semiconductor layers and embedded in the gate structure. The dielectric fin structure has a first width along the direction smaller than a second width of the isolation structure along the direction. The isolation structure includes a second dielectric layer extending through the gate structure and the first dielectric layer, and a third dielectric layer extending through the first dielectric layer and disposed on a bottom surface of the gate structure and a sidewall of the first dielectric layer.

Abstract: A semiconductor device includes a source/drain feature over a semiconductor substrate, channel layers over the semiconductor substrate and connected to the source/drain feature, a gate portion between vertically adjacent channel layers, and an inner spacer between the source/drain feature and the gate portion and between adjacent channel layers. The semiconductor device further includes an air gap between the inner spacer and the source/drain feature.

Abstract: A computer-implemented method for motion compensation of medical imaging data includes estimating, via a processor, non-periodic motion of a myocardium of a heart due to respiration and/or other body movements throughout a positron emission tomography (PET) scan based on list-mode emission data acquired during the PET scan of the heart of a subject. The method also includes performing, via the processor, event-by-event motion-corrected list-mode reconstruction on the list-mode emission data to generate cardiac images with the non-periodic motion removed.

Abstract: A medical image (24) with contrast of one imaging modality is emulated from images with other contrast characteristics. A plurality of input images (20) of a region of interest with the other contrast characteristics is received. A transform (22) which was generated by machine learning is applied to the corresponding plurality of input images (20) to generate a scalar value (38) or a vector value (38?) representing the corresponding voxel (36?) of the medical image (24) to be evaluated.

Abstract: A medical image (24) with contrast of one imaging modality is emulated from images with other contrast characteristics. A plurality of input images (20) of a region of interest with the other contrast characteristics is received. A transform (22) which was generated by machine learning is applied to the corresponding plurality of input images (20) to generate a scalar value (38) or a vector value (38?) representing the corresponding voxel (36?) of the medical image (24) to be evaluated.