89results about How to "Rich expression" patented technology

A strained-rough-voice conversion unit (10) is included in a voice conversion device that can generate a “strained rough” voice produced in a part of a speech when speaking forcefully with excitement, nervousness, anger, or emphasis and thereby richly express vocal expression such as anger, excitement, or an animated or lively way of speaking, using voice quality change. The strained-rough-voice conversion unit (10) includes: a strained phoneme position designation unit (11) designating a phoneme to be uttered as a “strained rough” voice in a speech; and an amplitude modulation unit (14) performing modulation including periodic amplitude fluctuation on a speech waveform. The amplitude modulation unit (14) generates, according to the designation of the strained phoneme position designation unit (11), the “strained rough” voice by performing the modulation including periodic amplitude fluctuation on the part to be uttered as the “strained rough” voice, in order to generate a speech having realistic and rich expression uttering forcefully with excitement, nervousness, anger, or emphasis.

The embodiment of the invention provides a video barrage displaying method and device. The method includes the following steps: receiving a spoiler shielding request sent by a first terminal, whereinthe spoiler shielding request includes a video identifier of a target video that is currently played; acquiring all barrage information of the corresponding target video according to the video identifier; matching all the barrage information with corresponding scenario keywords in a scenario keyword library; marking a spoiler tag on the successfully matched barrage information; filtering the barrage information marked with the spoiler tag; and sending the filtered barrage information to the first terminal to ensure that the filtered barrage information is displayed when the first terminal plays the target video. Thereby, according to the embodiment, the display forms of the barrage information can be enriched, the barrage information with spoiler contents can also be shielded according touser needs, and thus the user experience can be enhanced.

The invention relates to a virtual article display method and device, a terminal and a storage medium, and belongs to the technical field of networks. The method comprises the steps that when a virtual article custom instruction is received, displaying a virtual article custom window on a live broadcast interface; obtaining custom content input by a first user in the virtual article custom window,wherein the custom content is used for limiting the form of a to-be-generated virtual article; generating a virtual article corresponding to the custom content at least based on the custom content; and displaying the virtual article on the live broadcast interface. According to the invention, the audience user can customize the virtual article for the anchor user, the demand of customizing the virtual article by the audience user can be met, more abundant expression modes are provided for the audience user, the virtual article customized by the audience user can be displayed on the live broadcast interface, and the display effect of the virtual article can be enriched.

The invention belongs to the field of behavior recognition, particularly relates to a behavior feature extraction method, system based on space-time frequency domain hybrid learning, and a device, andaims to solve the problem of low skeleton behavior feature extraction precision. The method comprises the steps of obtaining a video behavior sequence based on a skeleton, and extracting a time-spacedomain behavior feature map through converting a network; inputting the time-space domain behavior feature map into a frequency domain attention network, performing frequency selection, inverting toa time-space domain, and adding the obtained behavior feature map to the time-space domain behavior feature map; synchronously performing local and non-local reasoning, and performing high-level localreasoning; and globally pooling the time-space domain behavior feature map obtained through reasoning to obtain the behavior feature vector of the video behavior sequence. The method can be applied to behavior classification, behavior detection and the like. According to the method, an effective frequency mode is adaptively selected in a frequency domain, a network with local affinity fields andnon-local affinity fields is adopted in a time-space domain for space-time reasoning, local details and non-local semantic information can be synchronously mined, and therefore the behavior recognition precision is effectively improved.

A festive atmosphere decorative lighting device which includes a lamp base, an LED bulb set inside the lamp base and an electrical switch for the bulb. A spinning assembly housed in the lamp base has an output shaft which is connected to an inner cover for spinning the inner cover. The inner cover fits over the LED bulb and has a cutout pattern of holes. The lamp base is also fitted with a transparent lamp shade which is superimposed above the inner cover. When the LED bulb is switched on and the spinning configuration spins the inner cover, light emitted by the LED bulb passes through the holes in the inner cover and presents a flowing vivid silhouette, as observed by a viewer.

The embodiment of the invention provides a base station data modeling method. The method comprises the steps that base station information of multiple base stations is acquired, wherein the base station information contains station IDs as well as longitudes and latitudes of the base stations; according to the base station information, the base stations are displayed in a geographic information system (GIS); adjacent region attributes of the station ID of the ith base station in the GIS are determined according to the station ID of the ith base station, wherein the ith base station is any one of the base stations, and the adjacent region attributes of the ith base station include Thiessen polygon data and adjacent region polygon data of the ith base station; according to the adjacent region attributes, the spatial rationality of the ith base station is analyzed. The embodiment of the invention further provides a terminal. When the method is adopted to analyze the rationality of the base stations, more abundant data structures and expression modes can be provided.

The invention relates to a three-dimensional imaging device capable of regulating shooting angles of two shooting lens. By a driving device in the three-dimensional imaging device, the shooting angles of two shooting lens are regulated to be crossed in front or behind a shot object, so that a shot three-dimensional picture generates a foreground or background three-dimensional effect, the effect of the three-dimensional picture can be more diversified and the manifestation mode of the three-dimensional picture is richer.

The invention discloses an image description generation method based on an external triple and an abstract relation. The method comprises the following steps: firstly, extracting a triple in an image description text, constructing an external relation library, and carrying out feature coding on the triple; and clustering the triples of which the text similarity is higher than a threshold value into one class. Meanwhile, the model performs target detection on the image to obtain a target visual feature set and a target category set; and querying a triple of which the target is similar to the target category in an external relation library according to the text similarity. The model predicts the target, the attribute and the relation of the image by using the visual features of the target to generate a scene graph; and fusing visual features and text features by using a convolutional neural network, and carrying out feature coding on targets, attributes and relationships. And finally, fusing target, attribute and relation coding features of the scene graph and coding features of the similarity relation and the abstract relation, and inputting the fused features into a double-layer LSTM sequence generation model to obtain final image description. According to the invention, the expression of the model generation description is richer.

The invention discloses a high-precision universal three-dimensional point cloud recognition method. The method comprises the following steps: performing nonlinear mapping on feature information of a point cloud to obtain representations in different transformation spaces; extracting local coding features from the obtained representations by using a graph attention network, obtaining global coding features through global space fusion, and obtaining fused point cloud features from the local coding features and the global coding features through a gated feature aggregation mechanism; sending the fused point cloud features into a classifier to obtain a prediction result; and constructing an overall prediction loss function, and adding independence discrimination loss for network training and optimization, and storing appropriate model parameters at the same time. According to the method, point cloud features are effectively extracted by using a graph attention convolutional neural network in each transformation, efficient fusion is performed through a gating mechanism, a Hilbert-Schmidt independence index is introduced to measure the similarity of the point cloud features, and information redundancy is reduced by minimizing the feature similarity, so that richer feature representation can be obtained.

The embodiment of the invention provides a data processing method and device, electronic equipment and a computer storage medium, and the method comprises the steps: determining that a triggering instruction for carrying out the image rendering of displayed to-be-processed data is received; obtaining attribute information of the to-be-processed data according to the trigger instruction; and obtaining a dynamic image corresponding to the attribute information, and displaying the dynamic image on a display interface of the to-be-processed data. Through the embodiment of the invention, the expression mode of the interactive scene is enriched, and the corresponding scene information can be effectively expressed.

The invention relates to a benign and malignant ulcer recognition method, which comprises the following steps: acquiring sample data, each piece of sample data comprising a plurality of pictures, and the sample data comprising unlabeled sample data and labeled sample data; carrying out coding pre-training on the encoder by using the unlabeled sample data, and obtaining a pre-trained picture to represent the encoder; inputting labeled sample data into the picture representation encoder, and outputting the feature representation of each picture; performing feature fusion on the feature representation of each picture to obtain a feature ultimate representation of the picture; and performing category prediction based on the feature ultimate representation of the picture. According to the method, the encoder is adopted to pre-train the unlabeled sample data, and then the labeled sample data is used for classification prediction, so that the workload of data labeling can be remarkably reduced while more comprehensive picture representation can be obtained, and the precision and efficiency of picture recognition are greatly improved.