摘要:
A method of generating a temporal saliency map is disclosed. In a particular embodiment, the method includes receiving an object bounding box from an object tracker. The method includes cropping a video frame based at least in part on the object bounding box to generate a cropped image. The method further includes performing spatial dual segmentation on the cropped image to generate an initial mask and performing temporal mask refinement on the initial mask to generate a refined mask. The method also includes generating a temporal saliency map based at least in part on the refined mask.
摘要:
A method of generating a temporal saliency map is disclosed. In a particular embodiment, the method includes receiving an object bounding box from an object tracker. The method includes cropping a video frame based at least in part on the object bounding box to generate a cropped image. The method further includes performing spatial dual segmentation on the cropped image to generate an initial mask and performing temporal mask refinement on the initial mask to generate a refined mask. The method also includes generating a temporal saliency map based at least in part on the refined mask.
摘要:
A system and method of video processing are disclosed. In a particular implementation, a device includes a processor configured to generate index data for video content. The index data includes a summary frame and metadata. The summary frame is associated with a portion of the video content and illustrates multiple representations of an object included in the portion of the video content. The metadata includes marker data that indicates a playback position of the video content. The playback position is associated with the summary frame. The device also includes a memory configured to store the index data.
摘要:
In some embodiments, a functional object-oriented network (FOON) is provided that includes a plurality of functional units each comprising two or more input object nodes that each identify an object and its state before a manipulation motion is performed, a motion node that identifies a manipulation motion that can be performed using the objects, and one or more output object nodes that each identify an object and its state after the manipulation motion has been performed using the objects. In some embodiment, a robot can used the FOON to determine the discrete actions that are required to perform a given task.
摘要:
A method, and product thereof, for modifying the surface of recycled sulfur-cured rubber via surface devulcanization that minimizes alteration of the network structure of the recycled sulfur-cured rubber, while generating strong tack between mixed recycled rubber and fresh rubber for optimal adhesion upon subsequent co-curing. In one embodiment, the surface of recycled rubber is devulcanized by treating recycled rubber with a thiol and a base dissolved in a solvent that does not effectively penetrate and only minimally swells vulcanized rubbers.
摘要:
In one or more embodiments, the present invention provides an way to modify the isoprene unit using an alder-ene reaction to form thermoset compounds comprising the resultant electron-deficient, readily crosslinkable polyisobutylene-based rubber that avoids the use of corrosive bromine or chlorine to make the activated butyl rubber, is easier to crosslink than the halobutyls, allows crosslinking with a simple organic base or a peroxide, and has mechanical properties as good as or better than sulfur crosslinked butyl rubbers.
摘要:
MEMS devices fabricated using inexpensive substrate materials such as paper or fabric, are provided. Using paper as a substrate, low cost, simple to prepare, lightweight, disposable piezoresistive sensors, including accelerometers are prepared. Signal-processing circuitry can also be patterned on the substrate material.The sensors can be utilized as two-dimensional sensors, or the paper substrate material can be folded to arrange the sensors in a three dimensional conformation. For example, three sensors can be patterned on a paper substrate and folded into a cube such that the three sensors are orthogonally positioned on the faces of a cube, permitting simultaneous measurement of accelerations along three orthogonal directions (x-y-z). These paper-based sensors can be mass produced by incorporating highly developed technologies for automatic paper cutting, folding, and screen-printing.Also provided are methods of modifying paper for use as a substrate material in MEMS devices.
摘要:
In general, optical communication systems and methods may generate higher-level nmQAM from nQAM signals using one or more delay line interferometers (DLIs) arranged in various configurations. The nQAM signals may be generated by a lower-level modulator, such as a BPSK modulator, QPSK modulator or lower-level QAM modulator, with binary driving signals. Different parameters of the DLIs, such as free spectral range (FSR), phase shift, and amplitude imbalance, may be selected to accomplish the desired higher-level nmQAM depending upon the nQAM signal.
摘要:
An integrated system for studying human grasping techniques includes a fingerprint sensor and a force sensor for simultaneously measuring the contact position and orientation of a fingertip together with the contact force and torque exerted by the fingertip during a grasp. The system uses the fingerprint image obtained from the fingerprint sensor to identify the contact position and orientation of the fingertip based on fingerprint features such as the core point and local ridge orientations. The contact position and orientation of the fingertip are represented in a fingertip coordinate system. The system has utility in human grasping studies to characterize the fingertip contact.
摘要:
Antigen binding proteins that bind to human IL-23 protein are provided. Nucleic acids encoding the antigen binding protein, vectors, and cells encoding the same as well as use of IL-23 antigen binding proteins for diagnostic and therapeutic purposes are also provided.