4results about How to "Reduce study load" patented technology

The invention discloses a double-arm robot control method based on single-arm trajectory prior guidance. The method comprises the following steps: acquiring a double-arm robot data set; skill primitives of different action types are obtained based on a single-arm action prediction model, the skill primitives of different action types are selected for a left arm and a right arm, and next action intentions of the left arm and the right arm are predicted respectively; constructing a double-arm diffusion generation model for establishing a mapping relation from initial action distribution to real double-arm coordination action distribution; training a single-arm action prediction model and a double-arm diffusion generation model based on the double-arm robot training data set; and acquiring motion prediction priori of the left arm and the right arm, constructing Gaussian distribution taking the single-arm priori as a mean value and a preset variance as a radius, sampling from the Gaussian distribution to obtain an initial motion vector, inputting the initial motion vector into a trained double-arm diffusion generation model, and de-noising the initial motion vector to generate double-arm motion. According to the invention, the independence of single-arm motion and the collaboration of double-arm control can be considered, and the double-arm motion control effect is optimized.

This application provides a spatial intelligent non-planar reflection generation method based on surface normal guidance. Multi-layer visual representations are extracted from the original RGB image of a scene containing a non-planar reflective medium. The non-planar reflective regions are identified by combining the brightness distribution and spatial symmetry of the original RGB image. The nonlinear deformation features of the texture in the non-planar reflective regions are analyzed, and pixel-level 3D surface normal vectors and confidence distributions are predicted using the multi-layer visual representations to construct a surface normal map. A gaze vector is determined based on a camera imaging model, and geometric optics mapping operations are performed on the gaze vector, surface normal map, and confidence distribution to construct a reflection offset field. Intermediate semantic feature representations corresponding to the non-planar reflective regions are extracted, and a pre-distorted reflection feature representation is generated based on the reflection offset field and the intermediate semantic feature representation. The pre-distorted reflection feature representation, multi-layer visual representation, and initial reflective region mask are injected into the generation model for image reconstruction, generating a non-planar reflective image.

The invention relates to the technical field of piano playing intelligent analysis, in particular to a motion capture method and system, a storage medium and a program product, and the method comprises the steps: obtaining the motion data of a first key; when it is recognized that the first key is in a pressed state from the first moment t1 to the second moment t2 according to the motion data, multiple groups of first hand images of the object from the third moment t3 to the fourth moment t4 are correspondingly collected; identifying a first hand executing a pressing action or having a pressing trend according to the first hand image; identifying a first finger executing a pressing action or having a pressing trend in the first hand; identifying a second key associated with the first finger; and judging whether the second key is matched with the identifier of the first key, and if so, correspondingly generating a motion capture result between the first moment and the second moment. According to the motion capture method provided by the invention, the fingering data of the object can be accurately acquired and matched to generate a reliable motion capture result.

The application provides a photovoltaic power short-term prediction method and device based on physical guidance knowledge distillation and a medium, and relates to the technical field of photovoltaic power prediction. The method comprises the following steps: obtaining an original data set related to photovoltaic power short-term prediction, and obtaining a plurality of physical characteristics through physical characteristic decoupling; constructing a theoretical power baseline of a to-be-predicted period, and calculating a residual sequence of historical power and the theoretical power baseline; constructing a multi-modal vector according to the plurality of physical characteristics and the residual sequence, inputting the multi-modal vector into a double-teacher model, and performing soft label integration on the output of the double-teacher model; taking the multi-modal vector as an input feature of a student model, taking a soft label and a pre-labeled real label as a supervision target, optimizing parameters of the student model, and outputting a predicted residual sequence; and reconstructing the predicted residual sequence and the theoretical power baseline to obtain photovoltaic predicted power. The application can improve the accuracy of photovoltaic power short-term prediction.