4results about How to "Correction bias" patented technology

The utility model discloses an external wheel forming die's correction mechanism, including base and support column, the support column symmetry fixed junction in base both sides to set up the correction subassembly in the base upper portion, the correction subassembly is used for heating correction to the external wheel, still include the rotating subassembly of setting up between the support column, the rotating subassembly is used for driving the external wheel and carries out the clamping rotation. The utility model belongs to external wheel forming die technical field, and specifically is the correction mechanism of automatic external wheel forming die of real -time monitoring the deviation situation of external wheel and automatic control correction accuracy.

The invention relates to the technical field of artificial intelligence and human-computer interaction, and discloses a rehabilitation training adjustment method and system based on multi-modal feature fusion and a storage medium, and the method comprises the steps: obtaining voice, face and physiological signals, constructing a multi-modal data stream, extracting features, and outputting a speech error classification identifier and a physiological signal quality index; splitting the facial action net displacement into healthy and affected sides, correcting affected side features based on healthy side features, and constructing visual representation; fusing each feature to output a comprehensive state vector; updating a dynamic baseline in a task gap, mapping a state vector, and screening to obtain an effective action set; calculating a composite reward and storing the composite reward in an experience playback pool to update the strategy network; and based on the action set, re-weighting the large model output probability, and generating a target interaction corpus. According to the method, the state sensing precision is improved by correcting the deviation of the affected side through the uninjured side, and a closed loop of self-adaptive adjustment of the rehabilitation difficulty and safe interaction text generation is realized by combining the dynamic baseline and penalty mechanism optimization reinforcement learning.

The utility model relates to the technical field of cylinder cover surface treatment, and disclose a kind of cylinder cover surface treatment frock, including base, the top of base is fixedly provided with two fixed plates, a fixed plate is fixedly provided with controller, another fixed plate is fixedly provided with observation glass window, the top of base is fixedly provided with positioning plate, the side of positioning plate is fixedly provided with first electric push rod, the side of first electric push rod is fixedly provided with clamping plate, the top of fixed plate is fixedly provided with processing frame, camera is fixedly provided on the processing frame, electric slide rail is fixedly provided in the width direction of processing frame.The cylinder cover surface treatment frock, by the cooperation of first electric push rod and clamping plate, can be accurately positioned and stably clamped according to different sizes of cylinder cover, avoid shaking or displacement in the processing process, ensure the uniformity and consistency of surface treatment effect, improve product quality.

This application discloses a radar echo extrapolation method based on temporal motion decomposition and dominant fusion, relating to the field of spatiotemporal sequence prediction technology. The method proposes an innovative network fusion paradigm, specifically including: acquiring implicit long-term / short-term temporal features and explicit global / transient motion features during processing using stacked hierarchical recurrent networks; then, innovatively constructing query vectors using temporal features and key-value vectors using motion features, and fusing them using a heterogeneous cross-attention mechanism to form a fusion architecture "dominated by an implicit state modeling network and referenced by an explicit motion modeling network." Based on this, through a series of collaborative designs such as directional decoupling, reference confidence estimation, and long-term feature constraints on short-term features, noise in long-term / short-term motion features is effectively suppressed, enhancing system stability. This method can improve the problems of blurred texture details and attenuation in high-value echo regions in predicted images.