Abstract: The techniques disclosed herein enable systems to perform repeatable and iterative load testing and performance benchmarking for artificial intelligence models deployed in a cloud computing environment. This is achieved by utilizing load profiles and representative workloads generated based on the load profiles to evaluate an artificial intelligence model under various workload contexts. The representative workload is then executed by the artificial intelligence model utilizing available computing infrastructure. Performance metrics are extracted from the execution and analyzed to provide insight into various performance dynamics such as the relationship between latency and data throughput. In addition, load profiles and input datasets are dynamically adjusted to evaluate different scenarios and use cases enabling the system to automatically test the artificial intelligence model across diverse applications.

Abstract: Pump assemblies for use with an IV set are described herein. A pump assembly includes a rotatable crank, a piston rod, and a diaphragm assembly. The crank includes a crank profile defining a plurality of lobes. The piston rod includes a first follower and a second follower spaced apart from the first follower. The rotation of the crank relative to the piston rod causes the piston rod to reciprocate relative to the crank. The diaphragm assembly includes a housing and a diaphragm member. The diaphragm member is coupled to the housing and cooperatively defines a volume in fluid communication with the IV set. The diaphragm member is movable between an extended position configured to draw in fluid from the IV set and a contracted position configured to expel fluid into the IV set. The diaphragm member is coupled to the piston rod to reciprocate the diaphragm member.

Abstract: There is proposed a mechanism for determining whether or not an imaging probe, such as an ultrasound imaging probe, is at a desired orientation and/or position with respect to an anatomical structure. Image data of the imaging probe is processed to generate a 3D landmark model that contains anatomical landmarks of the anatomical structure. The 3D landmark model is then processed to determine whether or not the imaging probe is at the desired orientation and/or position.

Abstract: An embodiment for creating an avatar of an online viewer during a live video broadcast is provided. The embodiment may include receiving a number of in-person viewers at an event. The embodiment may also include identifying a number of online viewers watching the event remotely. The embodiment may further include in response to determining the number of in-person viewers at the event is below a threshold, identifying a location of the in-person viewers at the event. The embodiment may also include analyzing reactions of the online viewers to scenarios of the event. The embodiment may further include creating an avatar for each of the online viewers. The embodiment may also include populating each empty seat at the event with the created avatar for each online viewer. The embodiment may further include displaying the created avatars occupying the empty seats to the online viewers.

Abstract: There is provided a method (40) of correcting blade pitch in a wind turbine (10) having a tower (12) and a plurality of rotor blades (18). The method (40) comprises: receiving (410) sensor output data from one or more wind turbine sensors (141), the sensor output data including data indicative of excitation of the tower (12) for a plurality of different pitch angles of a particular one of the blades (18); determining (420) a corrected pitch reference of the particular blade (18), the corrected pitch reference corresponding to a minimum tower excitation based on the received sensor output data; and, sending (430) the corrected pitch reference to a pitch actuator system (24) of the particular blade (18).

Abstract: Aspects of the present invention relate to a method for controlling one or more wind turbine generators. The method comprises: receiving a trigger signal, and implementing an active power boost mode during a predetermined time period in response to receiving the trigger signal. The active power boost mode comprises: disregarding further trigger signals; and generating one or more upper active power limits for each of the one or more wind turbine generators, wherein, during at least a first portion of the predetermined time period, the one or more upper active power limits are fixed at an active power boost level that is greater than a nominal active power level of the one or more wind turbine generators.

Abstract: There is proposed a mechanism for generating and displaying a 3D representation of an anatomical structure of an individual. Image data of the anatomical structure is obtained and processed to obtain 2D images or image sequences corresponding to predetermined views of the anatomical structure. Anatomical landmarks are identified in the 2D images or image sequences and used to determine a 3D landmark model of the anatomical 5structure. The 3D landmark model is used to render and display a 3D representation of the anatomical structure.

Abstract: In some embodiments, the present specification describes methods for displaying a three-dimensional image of an isolated threat object or region of interest with a single touch or click and providing spatial and contextual information relative to the object, while also executing a view dependent virtual cut-away or rendering occluding portions of the reconstructed image data as transparent. In some embodiments, the method includes allowing operators to associate audio comments with a scan image of an object. In some embodiments, the method also includes highlighting a plurality of voxels, which are indicative of at least one potential threat item, in a mask having a plurality of variable color intensities, where the intensities may be varied based on the potential threat items.

Abstract: A method and an electronic device (100) are disclosed for generating semantic representation of a document to determine data security risk associated with the document. The method includes receiving, by a document semantics controller (160) of the electronic device (100), a document in an electronic form and determining, by the document semantics controller (160), raw text. Further, the method includes generating, by the document semantics controller (160), a plurality of sentence blocks using the raw text and determining, by the document semantics controller (160), embeddings for the plurality of sentence blocks. Further, the method includes determining, by the document semantics controller (160), the semantic representation of the document based on the embeddings for each of the sentence blocks; and generating, by the document semantics controller (160), the semantic representation of the document to determine the data security risk associated with the document.

Abstract: Unoccluded visualization of Anatomy in Volumetric Data. Embodiments disclosed herein relate to medical imaging systems, and more particularly to medical visualization systems, which can be used for visualizing anatomy. Embodiments herein disclose methods and systems for enabling unoccluded visualization of anatomy of interest in volumetric data. Embodiments herein disclose 3D visualization techniques for visualization of deep seated anatomy with no user interactions, while preserving depth continuity with surrounding(s) and with minimal distortion for improved clinical assessment.

Abstract: A method and an electronic device (100) are disclosed for generating semantic representation of a document to determine data security risk associated with the document. The method includes receiving, by a document semantics controller (160) of the electronic device (100), a document in an electronic form and determining, by the document semantics controller (160), raw text. Further, the method includes generating, by the document semantics controller (160), a plurality of sentence blocks using the raw text and determining, by the document semantics controller (160), embeddings for the plurality of sentence blocks. Further, the method includes determining, by the document semantics controller (160), the semantic representation of the document based on the embeddings for each of the sentence blocks; and generating, by the document semantics controller (160), the semantic representation of the document to determine the data security risk associated with the document.

Abstract: A computer-implemented method for visualization of an elongated anatomical structure (20), for example of a fetal spine using ultrasound is provided.

Abstract: The invention provides a method for deriving a biometric parameter of a fetal heart. The method includes acquiring a plurality of ultrasound images of a region of interest, wherein the region of interest comprises a fetal heart and comparing the plurality of ultrasound images to a predefined clinical view. A group of ultrasound images related to the predefined clinical view are selected based on the comparison, wherein the group of ultrasound images represents at least one cardiac cycle. An anatomical landmark of the fetal heart is detected within an ultrasound image of the group of ultrasound images and the anatomical landmark of the fetal heart is detected or tracked across the group of ultrasound images. A biometric parameter of the fetal heart is then determined based on the detected or tracked anatomical landmark from one or more ultrasound images of the group of ultrasound images.

Abstract: In some embodiments, the present specification describes methods for displaying a three-dimensional image of an isolated threat object or region of interest with a single touch or click and providing spatial and contextual information relative to the object, while also executing a view dependent virtual cut-away or rendering occluding portions of the reconstructed image data as transparent. In some embodiments, the method includes allowing operators to associate audio comments with a scan image of an object. In some embodiments, the method also includes highlighting a plurality of voxels, which are indicative of at least one potential threat item, in a mask having a plurality of variable color intensities, where the intensities may be varied based on the potential threat items.

Abstract: Methods for improving a toughness and a strength of a stainless steel material are described herein. For example, a high strength stainless steel material can comprise at least 11 wt. % Cr, between 0.01 wt. % and 1.0 wt. % Ni, more 0 wt. % Mo, more than 0 wt. % W, more than 0 wt. % Ti, more than 0 wt. % Nb, and more than 0 wt. % V. In some examples, the high strength stainless steel material can be heat treated with at least one quench treatment and at least one tempering heat treatment. In some examples, the high strength stainless steel material can comprise between 0.01 wt. % and 0.5 wt. % Ni, no more than 0.25 wt. % Mo, no more than 0.1 wt. % W, no more than 0.1 wt. % Ti, no more than 0.1 wt. % Nb, and no more than 0.1 wt. % V.

Abstract: The invention relates to a computer-implemented method for automatically detecting anatomical structures (3) in a medical image (1) of a subject, the method comprising applying an object detector function (4) to the medical image, wherein the object detector function performs the steps of: (A) applying a first neural network (40) to the medical image, wherein the first neural network is trained to detect a first plurality of classes of larger-sized anatomical structures (3a), thereby generating as output the coordinates of at least one first bounding box (51) and the confidence score of it containing a larger-sized anatomical structure; (B) cropping (42) the medical image to the first bounding box, thereby generating a cropped image (11) containing the image content within the first bounding box (51); and (C) applying a second neural network (44) to the cropped medical image, wherein the second neural network is trained to detect at least one second class of smaller-sized anatomical structures (3b), thereby g

Abstract: The invention relates to a method for determining power set points of a power plant comprising a plurality of power generating units with at least one wind turbine generator. The determination of the power set points are based on a prioritization of the power generating units, where the power generating units are prioritized with respect to individual power levels or fatigue levels obtained for the power generating units. The prioritization is adjusted so that the adjusted prioritization depends both on the power and fatigue levels.

Abstract: There is provided a method (40) of correcting blade pitch in a wind turbine (10) having a tower (12) and a plurality of rotor blades (18). The method (40) comprises: receiving (410) sensor output data from one or more wind turbine sensors (141), the sensor output data including data indicative of excitation of the tower (12) for a plurality of different pitch angles of a particular one of the blades (18); determining (420) a corrected pitch reference of the particular blade (18), the corrected pitch reference corresponding to a minimum tower excitation based on the received sensor output data; and, sending (430) the corrected pitch reference to a pitch actuator system (24) of the particular blade (18).

Abstract: Provided is a junction, a method for forming a junction, and a well system including a junction. The junction, in one aspect, includes a first threaded member, the first member formed of a first material, and a second threaded member threaded with the first threaded member, the second threaded member formed of a second material. In one or more aspects, the first and second threaded members define an overlapping space. In one or more other aspects, an expandable metal joint is located in at least a portion of the overlapping space, the expandable metal joint comprising a metal configured to expand in response to hydrolysis.

Abstract: An embodiment for creating an avatar of an online viewer during a live video broadcast is provided. The embodiment may include receiving a number of in-person viewers at an event. The embodiment may also include identifying a number of online viewers watching the event remotely. The embodiment may further include in response to determining the number of in-person viewers at the event is below a threshold, identifying a location of the in-person viewers at the event. The embodiment may also include analyzing reactions of the online viewers to scenarios of the event. The embodiment may further include creating an avatar for each of the online viewers. The embodiment may also include populating each empty seat at the event with the created avatar for each online viewer. The embodiment may further include displaying the created avatars occupying the empty seats to the online viewers.