Abstract: Provided are a training method and apparatus for a target detection model, a device and a storage medium. The training method is described below. A feature map of a sample image is processed through a classification network of an initial model and a heat map and a classification prediction result of the feature map are obtained, a classification loss value is determined according to the classification prediction result and classification supervision data of the sample image, and a category probability of pixels in the feature map is determined according to the heat map of the feature map and a probability distribution map of the feature map is obtained; the feature map is processed through a regression network of the initial model and a regression prediction result is obtained, and a regression loss value is determined.

Abstract: An electromagnetic and triboelectric hybrid energy collector for low-frequency movement is provided. The electromagnetic and triboelectric hybrid energy collector comprises an electromagnetic module, triboelectric modules and spring mass modules. The electromagnetic module includes a shell, a magnet frame, a magnet array, coil arrays, coil frames and coil magnetic conductive columns. The spring mass modules include respective springs, respective mass blocks and respective end caps. The triboelectric modules include multiple nano-friction power generation units which are connected in sequence. When the energy collector is excited to move a magnet assembly. The electromagnetic module generates voltage according to the Faraday's law of electromagnetic induction. The movement of the magnet assembly generated by external excitation may also cause contact and separation between triboelectric units to generate voltage.

Abstract: Aspects of the invention include receiving an original work written in a first style to be transformed into a derivate work written in a second style. The transforming includes accessing a similarity data structure organized to include a phrase-similarity vector space of phrase vectors representing degrees of similarity between phrases in a corpus of works written in the second style. An original sentence of the original work that includes one or more original phrases is obtained. The original phrases are located in the phrase-similarity vector space. For each of the original phrases, a similar phrase from the phrase-similarity vector space is selected based on the selected similar phrase having a phrase-similarity index corresponding to the original phrase that is greater than a predetermined phrase-similarity index threshold. The derivative work is created by replacing the original phrases in the original work with the selected similar phrases. The derivative work is output.

Abstract: Embodiments of the present invention provide a carrier resource configuration method, an uplink information transmission method, a base station, and a user equipment. The configuration method includes: determining, by the base station, a serving cell group to which a serving cell of the UE belongs; and sending, by the base station, a belonging relationship between a serving cell and a serving cell group to the UE, so that the UE performs uplink transmission according to the belonging relationship. The technical solutions of the present invention solve a problem that a UE cannot be scheduled correctly because a base station cannot acquire UCI of the UE in time.

Abstract: The disclosure provides a method for recognizing an image, an apparatus for recognizing an image, an electronic device and a storage medium. An image to be processed is obtained. The number of first channels of the image is greater than the number of second channels of a red-green-blue (RGB) image. For each pixel of the image, a semantic type of the pixel is determined based on a value of the pixel on each channel. A recognition result of the image is generated based on the image and the semantic type of each pixel.

Abstract: An object detection model training method, object detection method and related apparatus, relate to the field of artificial intelligence technologies such as computer vision, deep learning. An implementation includes: obtaining training sample data including a first remote sensing image and position annotation information of an anchor box of a subject to be detected in the first remote sensing image, where the position annotation information includes angle information of the anchor box relative to a preset direction; obtaining an object feature map of the first remote sensing image based on an object detection model, performing object detection on the subject to be detected based on the object feature map to obtain an object bounding box, and determining loss information between the anchor box and the object bounding box based on the angle information; updating a parameter of the object detection model based on the loss information.

Abstract: The present invention discloses a quasi-zero-stiffness (QZS) based six-degree-of-freedom (6-DOF) absolute displacement and attitude measurement device. A lower end coil and an upper end coil are respectively charged with currents in the opposite directions; The electromagnetic field and the magnetic fields of an upper magnet and a lower magnet per se are mutually acted to produce an electromagnetic stiffness opposite to the stiffness of a spring. The stiffness of the whole leg is close to zero stiffness. When the to-be-measured platform generates space motion, the reference platform is in the stationary state. At this point, the deformation amounts of the six legs can be measured by laser displacement sensors. The six deformation amounts are respectively inputted into the displacement and attitude resolver, and by forward kinematic solution of the 6-DOF device, the displacement and the attitude of the to-be-measured platform can be obtained.

Abstract: The present disclosure provides a single-degree-of-freedom (SDOF) magnetic damping shock absorber based on an eddy current effect, comprising a lower plate, a ring-shaped magnet a, a ring-shaped magnet b, an aluminum cylinder, a bottom copper sheet, a copper sheet, a top copper sheet, a bearing seat, a linear bearing, a bearing end cap, a load, a piston shaft, a stepped shaft, a fixed collar, a coil spring, a lower clamping shaft, and fixing screws. When the shock absorber is working, the ring-shaped magnet a keeps stationary at the lower end and the ring-shaped magnet b reciprocates in the vertical direction. Both magnets are arranged in a mutual attraction manner. Under the action of a time-varying electromagnetic field generated by the relative movement of the ring-shaped magnet b, the copper sheet arranged between the two ring-shaped magnets generates eddy current damping. The movement of the ring-shaped magnet b is inhibited.

Abstract: An object area measurement method and an apparatus are provided, relating to the computer vision and deep learning technology. The method includes acquiring an original image with a spatial resolution, the original image including a target object; acquiring an object identification model including at least two sets of classification models; generating one or more original image blocks based on the original image; performing operations on each original image block: scaling each original image block at at least two scaling levels to obtain scaled image blocks with at least two sizes, the scaled image blocks respectively corresponding to the at least two sets of classification models, and inputting the scaled image blocks into the object identification model to obtain an identification result of the target object; and determining an area of the target object based on the respective identification results of the one or more original image blocks and the spatial resolution.

Abstract: The present disclosure discloses a method and apparatus for generating an image, a device, a storage medium and a program product, relates to the field of artificial intelligence, and particularly to computer vision and deep learning technologies, and may be applied in smart cloud and power grid inspection scenarios. A particular implementation of the method comprises: acquiring an original insulator image; performing an image transformation on the original insulator image to obtain a composite insulator image; and inputting the original insulator image and the composite insulator image into a pre-trained generative adversarial network to generate a target insulator image. According to the implementation, the image transformation is performed on the original insulator image, and then, massive target insulator images are generated through the generative adversarial network.

Abstract: A wheel-legged amphibious mobile robot with a variable attack angle, which belongs to the technical field of robot structure technology. The robot includes three parts: motion unit, body trunk and power unit. As a key structure, the motion unit mainly includes a moving mechanism, a wheel assembly, a telescopic mechanism and a transmission device. The robot drives the telescopic mechanism to reciprocate linearly through a gear and rack set, and pushes “legs” to expand and retract, so as to realize a mutual switching between a wheeled mode and a gait mode. Under transmission of bevel gear set, the blades can rotate at any same angle at the same time, to change the attack angle and realize the steering. The robot provided by the present disclosure can effectively adapt to a complex and harsh amphibious environment, and meet a series of operation requirements such as rapid movement, obstacle climbing, underwater steering.

Abstract: A method for qualitative identification and quantitative determination of caffeine in a drug. In the method, terahertz absorption coefficient spectra of drugs with different concentrations of caffeine are measured, from which the frequency points, amplitudes and peak areas of characteristic peaks of the drugs with different concentrations of caffeine are obtained as characteristic quantities. Concentration gradients are established between the concentrations and the characteristic quantities, respectively. The characteristic quantities are imported to the SVR model to establish a training set and a test set. Finally, the qualitative identification and quantitative analysis of caffeine in unknown drugs are achieved.

Abstract: The present disclosure discloses a method and apparatus for training a model, a method and apparatus for predicting a mineral, a device and a storage medium, and relates to the fields of computer vision and deep learning technologies. An implementation of the method may include: acquiring a target hyperspectral image of a target area, the target hyperspectral image including at least one pixel point annotated with a mineral category; determining a mask image corresponding to the target hyperspectral image; determining a sample hyperspectral image according to the target hyperspectral image and the mask image; determining an annotation vector of each pixel point according to the at least one pixel point annotated with the mineral category; and training a model according to the sample hyperspectral image and the annotation vector of the each pixel point.

Abstract: Provided are a method and apparatus for evaluating image relative definition, a device and a medium, relating to technologies such as computer vision, deep learning and intelligent medical. A specific implementation solution is: extracting a multi-scale feature of each image in an image set, where the multi-scale feature is used for representing definition features of objects having different sizes in an image; and scoring relative definition of each image in the image set according to the multi-scale feature by using a relative definition scoring model pre-trained, where the purpose for training the relative definition scoring model is to learn a feature related to image definition in the multi-scale feature.

Abstract: A super-tough steel and a production method thereof. The present invention belongs to the field of steels and preparation thereof, and particularly relates to a super-tough steel for cryogenic services, and a production method thereof. The super-tough steel comprises the following elements in percentage by weight: 0.10% to 0.15% of C, and 29.5% to 31.5% of Mn, with the balance being Fe and inevitable impurities. The production method comprises the following steps: A1, smelting with argon protection, and electroslag remelting; A2, hot rolling or hot forging; A, annealing at 900° C. to 1100° C. for 1 hour, and quenching; and A4, cold rolling, annealing of the cold-rolled plate at 700° C.-1200° C. for 1 hour, and quenching after annealing. The steel is simple in composition and does not contain precious metals; and the steel has an average grain size of less than 30 microns, and is non-magnetic with complete face-centered cubic structure.

Abstract: A header model for instance segmentation includes a target box branch having a first branch and a second branch, where the first branch is configured to process an inputted first feature map to obtain class information and confidence of a target box, and the second branch is configured to process the first feature map to obtain location information of the target box. The header model also includes a mask branch configured to process an inputted second feature map to obtain mask information, wherein the second feature map is a feature map outputted by an ROI extraction module, and the first feature map is a feature map resulting from a pooling performed on the second feature map.

Abstract: The present disclosure provides a terahertz biological detection method and a device comprising the same, in which biological samples with different molecular formulas have their characteristic peaks in the terahertz band. In the process of sweeping frequency with terahertz, when the terahertz frequency point corresponds to the characteristic peak frequency of the substance to be detected, resonant absorption occurs, and the transmitted/reflected terahertz wave electric field intensity will suddenly decrease. In the electromagnetically induced transparency spectrum corresponding to the Rydberg quantum state, the signal splitting amplitude is significantly reduced. Therefore, the characteristic peak frequency and specific content of the substance to be detected can be accurately determined by comparing the dependence of the Rydberg quantum state with the additional terahertz electric field intensity on the excitation energy level.

Abstract: Provided is a method for identifying authenticity and origin of Panax quinquefolius based on terahertz spectroscopy. The time domain spectral information of the Panax quinquefolius sample is obtained, and converted into the frequency domain spectral information by Fourier transform to calculate a terahertz absorption spectrum. The identification is performed by observing whether there are characteristic absorption peaks of pseudoginsenoside F11 in the terahertz absorption spectrum.