Abstract: A device inputs a raw image into a machine learning model, and receives a predicted image comprising a plurality of masks as output from the machine learning model, each mask of the plurality of masks surrounding a plurality of pixels predicted to correspond to a given classification predicted by the machine learning model. The device inputs the raw image and the predicted image into a refinement model, and receives, as output from the refinement model, revised boundaries for each mask of the plurality of masks, the revised boundaries generated by fusing each mask with a boundary determined from the raw image by using a segmentation model. The device outputs classifications for each pixel of the raw image based on the revised boundaries.

Abstract: A harmonic reducer includes a power shaft, a wave generator mounted on the power shaft, a circular spline including internal teeth, a flex spline including external teeth and configured to receive the wave generator to cause the external teeth to mesh with the internal teeth progressively as the wave generator rotates. The harmonic reducer further includes a bearing system configured to support the power shaft, and the bearing system includes at least one first bearing and a second bearing located at a first side of the wave generator.

Abstract: A fixing module and a processing device are provided. The fixing module includes a first clamping component, a second clamping component, a fixing component, a rotation component, and an elastic component. The first clamping component is configured to clamp an object. The second clamping component corresponds to the first clamping component and is configured to clamp the object. The fixing component corresponds to the second clamping component. The rotation component corresponds to the second clamping component. The elastic component is connected to the second clamping component and the fixing component, and a portion of the elastic component is arranged on the rotation component. The fixing module is configured to pull the second clamping component through the elastic component, such that a tension force is applied to the object by the second clamping component.

Abstract: The present disclosure relates to a probe adapter and a measurement system, including: a first interface, configured to be electrically connected to a probe; a second interface, configured to be electrically connected to a measuring instrument; signal transmission lines, including first and second signal transmission lines, and a line selection switch, where the first and second signal transmission lines are both located between the first interface and the second interface, and the line selection switch is configured to select the first signal transmission line or the second signal transmission line to conduct the first interface and the second interface; and a signal conditioning circuit, located on the second signal transmission line, and configured to adjust an impedance of a measured signal to match an input impedance of the measuring instrument, and/or adjust an amplitude of the measured signal to match a measurement range of the measuring instrument.

Abstract: A piezoionic device for providing an electrical signal, wherein the device comprising: (i) a matrix comprising a plurality of mobile cations and anions, (ii) two or more electrodes in contact with the matrix, configured to provide the electrical signal in response to an external pressure applied to the matrix depending on a mobility difference between the cations and the anions. The piezoionic device comprises a selective ion mobility amplification mechanism, for selectively enhancing or hindering diffusivity of either the cations or the anions, thereby increasing the mobility difference which amplifies the electrical signal.

Abstract: A system or a method for fast training an artificial intelligence (AI) model. The system accesses a training dataset, divides the training dataset into a training subset and a holdout test subset, and trains the AI model using the training subset. For each of multiple training epochs, the system checkpoints while training the AI model. Checkpointing includes applying the trained AI model to the training subset to determine a train loss value, applying the trained AI model to the holdout test subset to determine a test loss value, and determining whether both train loss values and test loss values converge at a particular epoch. Responsive to determining that convergence occurred at the particular epoch, the system selects the AI model trained at the particular epoch.

Abstract: Disclosed is a loudspeaker structure including a housing, a magnetic element, a conductive ring and a first vibration assembly. The housing has an accommodation groove and a first sound cavity, the accommodation groove communicates with the first sound cavity, and the first sound cavity has a first opening. The magnetic element is disposed in the accommodation groove. The conductive ring is wound around a peripheral side of the magnetic element, and a current in the conductive ring controls a magnetic field of the magnetic element. The first vibration assembly has a first diaphragm and a first vibration element, the first diaphragm covers the first opening, the first vibration element is disposed on the first diaphragm, and the first vibration element is attracted or repelled by the magnetic field of the magnetic element to drive the first diaphragm to vibrate to produce sound.

Abstract: A device generates a training set of images by, for each image of a plurality of training images, receiving user input of a set of labels for a portion of the image, the portion less than an entirety of the image, the set of labels comprising classifications of individual pixels within the image, and automatically applying a label of unknown to a remainder of the image that excludes the portion of the image. The device inputs an unlabeled image into a machine learning model, the machine learning model trained using the training set, and receives, as output from the machine learning model, predicted classifications for each pixel of the image.

Abstract: The application provides a display panel motherboard and manufacturing method thereof. The display panel motherboard includes display panel forming areas distributed in an array and a test area adjacent to the display panel forming areas. The display panel motherboard includes an array mother substrate; a light-emitting layer, disposed on a surface of a side of the array mother substrate and located in the display panel forming areas; and a test component, disposed on the surface of the array mother substrate and located in the test area, the test component including a test block including a fluorescent layer, disposed on the surface of the array mother substrate and emitting light upon being irradiated by activation light; a target test layer, disposed on the side of the fluorescent layer facing away from the surface; and a positioning reference portion, disposed at a peripheral side of the target test layer.

Abstract: A rubber reinforcing filler and a method for preparing rubber reinforcing filler by activation and modification of coal gangue are provided. The method includes the following steps: providing coal gangue raw material; crushing coal gangue raw materials; the crushed coal gangue raw materials are ultrafine grinding to obtain coal gangue particles; coal gangue particles were obtained by air flow classification. Activation of coal gangue particles; the activated gangue particles were surface modified and ground to obtain the modified gangue particles. The invention relates to the field of comprehensive utilization of coal gangue and the technical field of rubber filler. The method is simple and effective, which can complete the blending modification of coupling agent and activated coal gangue particles for surface modification at low cost, so as to avoid the waste of materials in the traditional modification process and realize the preparation of high quality rubber reinforcing filler.

Abstract: Provide crystal forms of 5-fluoro-1-(4-fluoro-3-(4-(pyrimidin-2-yl) piperazine-1-carbonyl) benzyl) quinazoline-2, 4 (1H,3H)-dione and preparation method thereof.

Abstract: A composition comprising catalase, a preparation method and a use thereof, and a method for killing tumor cells are provided. The composition comprises a radionuclide labeled to a biomacromolecule, a soluble alginate and catalase. The composition can be injected into the tumor through an interventional treatment. A gel is formed when an alginate ion in the composition enters the tumor and encounters a calcium ion, such that the radionuclide and the catalase are uniformly confined in the tumor. The composition comprising catalase utilizes catalase to decompose dissolved oxygen generated from hydrogen peroxide in the tumor to advance the hypoxic state of the tumor cells, and the tumor cells are killed with radiation after the hypoxic state thereof has been advanced, and thus the invention has good prospects for applications in cancer therapy.

Abstract: Disclosed are cartridge components, cartridges, systems, and methods for isolating analytes from biological samples. In various aspects, the cartridge components, cartridges, systems, and methods may allow for a rapid procedure that requires a minimal amount of material from complex fluids.

Abstract: The present disclosure discloses an application of Loigolactobacillus coryniformis subsp. coryniformis in the efficient degradation of aflatoxin. The method is as follows: Loigolactobacillus coryniformis subsp. coryniformis 523L5 is inoculated into LB broth containing Aflatoxin B1, and Aflatoxin B1 is degraded by dark reaction at 25-55° C. and pH 3.0-9.0. The Loigolactobacillus coryniformis subsp. coryniformis 523L5 has good growth performance, high acid production ability, strong salt tolerance and acid tolerance, and may tolerate about 7% salinity. It may grow under the condition of pH 4.5 and has good fermentation performance. The fermentation broth can efficiently degrade Aflatoxin B1 under acidic (pH 5.0-7.0) and high temperature (50-55° C.) conditions, and the highest degradation rate can reach 88.13% at 40° C. and pH 7.0.

Abstract: A nonlinear oxygen-enriched injection method based on chaotic mapping and electronic device is disclosed, including: obtaining a chaotic gas injection volume corresponding to a current speed change period according to a chaotic mapping value corresponding to the current speed change period and a peak gas injection volume in an oxygen-enriched injection process; determining a rotational speed of a fan blade in a fan component corresponding to the current speed change period according to the chaotic gas injection volume; updating a rotational speed of a direct current (DC) motor in the fan component corresponding to the current speed change period according to the rotational speed of the fan blade in the fan component, and driving the fan blade to rotate according to an updated rotational speed of the DC motor, so as to update an air output of the fan component. The above operations are repeated until a last stage.

Abstract: A system or method for identifying mislabels in a training dataset. The labeled dataset includes multiple labeled samples, each of which is labeled with a ground-truth label. The labeled samples are divided into multiple training subset and hold-out test subset. For each training subset and corresponding hold-out test subset, a machine learning model is trained using the training subset, and the trained machine learning model is applied to the hold-out test subset to generate prediction labels for samples in the corresponding hold-out test subset. Each prediction label has a confidence score indicating a likelihood of the prediction label is correct. A pair of prediction label and ground truth label corresponding to a same sample is compared to determine whether there is a candidate mislabel. The system determines whether the candidate mislabel is a mislabel based in part on a confidence score of the prediction label.

Abstract: A least-squares reverse time migration method, system, terminal, and non-transitory computer-readable storage medium.

Abstract: A method for preparing a nano metal paste, including the following steps. (S1) An organic metallic salt and an organic reducing agent are added into water or an organic solvent to obtain a first mixture. (S2) The first mixture is reacted under heating to obtain a second mixture containing metal nanoparticles. (S3) The second mixture is concentrated to obtain a nano metal paste containing 20-95% by weight of the metal nanoparticles. A conductive paste and A conductive ink prepared by using the nano metal paste as filler are also provided.