Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for generating a three-dimensional (3D) object model. The method includes generating a first 3D object model based on multiple two-dimensional (2D) images of an object in different views. The method further includes acquiring metadata related to the first 3D object model by searching for information related to the object in at least one of a database and the Internet. The method further includes generating a second 3D object model by combining the first 3D object model and the metadata. The method for generating a 3D object model according to the present disclosure can automatically generate customizable and editable 3D model metadata, thereby significantly reducing labor, saving costs, improving efficiency, and improving user experience.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for processing point clouds. The method includes performing upsampling on a first feature of a first point cloud of a target object. The method further includes determining a reference feature of a second point cloud of a reference object. The method further includes determining a second feature based on the first feature subjected to upsampling and the reference feature. The method further includes generating a third point cloud of the target object based on the second feature and the second point cloud of the reference object, where the third point cloud has a larger number of points than the first point cloud. Through embodiments of the present disclosure, a point cloud of the target object can be made denser, with increased accuracy, thereby providing a more detailed description of the target object.

Abstract: A method in an illustrative embodiment includes: obtaining a first point cloud based on an input point cloud, a point number of the first point cloud being greater than a point number of the input point cloud; obtaining a first group of point clouds based on the first point cloud, the first group of point clouds including a plurality of point clouds; obtaining a second group of point clouds based on the input point cloud and the first group of point clouds, the second group of point clouds including a plurality of point clouds; and obtaining a target point cloud based on the first point cloud and the second group of point clouds, a point number of the target point cloud being greater than the point number of the input point cloud.

Abstract: The present disclosure relates to a method, a device, and a computer program product for image processing. The method includes acquiring a plurality of images and a first video, wherein the plurality of images indicate visual information of a source object from a plurality of perspectives, and the first video indicates animation of a target object. The method further includes generating a three-dimensional model for the source object based on the plurality of images, and generating a plurality of animation models for the target object based on the first video. The method further includes fusing the three-dimensional model for the source object and the plurality of animation models for the target object to generate a second video for the source object, wherein in the second video, the target object in the first video is replaced with the source object.

Abstract: Embodiments of the present disclosure provide a method, an electronic device, and a computer program product for processing a watermark of a neural network model. The method includes: embedding a parameter component watermark into a first parameter of a neural network model to generate a second parameter of the neural network model; embedding an input component watermark into a first input to the neural network model to generate a second input to the neural network model; embedding a gradient component watermark into a first model gradient of the neural network model to generate a second model gradient of the neural network model; and training the neural network model based on the second parameter, the second input, and the second model gradient to generate a trained neural network model.

Abstract: Embodiments of the present disclosure relate to a method, a device, a medium, and a program product for training a question-answer system. The method includes: determining a distribution of hidden variables in a variational language model based on a query in a training data set. The method further includes: generating a plurality of answers for the query using the variational language model based on a plurality of hidden variables randomly sampled from the distribution. The method further includes: determining reward scores for the plurality of answers using a reward model. The method further includes: updating the variational language model based on the query and the best answer with the highest reward score among the plurality of answers.

Abstract: The present disclosure relates to a method, a device, and a computer program product for generating a super resolution image. The method includes: generating, by a first network and based on a first image of first resolution, a second image of first super resolution; determining, by the first network, a first residual image based on the first image and the second image; and generating, by a second network, a third image of second super resolution based on the first residual image and the second image, wherein the first super resolution is higher than the first resolution and the second super resolution is higher than the first super resolution. In this way, the data fidelity can be maintained and the signal-to-noise ratio can be reduced when generating high resolution images, thereby improving the image quality.

Abstract: Embodiments of the present disclosure provide a method, a device, and a computer program product for generating a super-resolution image model. The method includes acquiring a first image with a first resolution and a second image with a second resolution, the first image corresponding to the second image; generating a first super-resolution image with a first super resolution and a second super-resolution image with a second super resolution according to an initial super-resolution image model based on the first image; transforming the first super-resolution image into a first frequency-domain representation; transforming the second super-resolution image into a second frequency-domain representation; and generating a trained super-resolution image model based on a loss between the first frequency-domain representation and the second frequency-domain representation and a reference frequency-domain representation of the second image.

Abstract: Embodiments of the present disclosure provide a method for determining a generative model. The method includes embedding a white box watermark and a black box watermark into a generative model. The black box watermark is first embedded into a probability density function of data abstractions in respective layers of the generative model. The method further includes embedding, after the embedding of the black box watermark is completed, the white box watermark into respective layers for outputs of the generative model. Model data is generated by the generative model based on predetermined triggering data. The predetermined triggering data includes a predetermined triggering text or a predetermined triggering image. An identity associated with the generative model is determined based on the model data. Advantageously, the illustrative method is capable of providing double-layer protection for a generative model by embedding two complementary and independent watermarks to resist white box and black box attacks.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for image processing. The method includes: obtaining an encoding feature of a reference image and an encoding feature of an input image of a first resolution, wherein the reference image has a resolution greater than the first resolution. The method further includes: obtaining high-frequency information and low-frequency information on the input image by interpolating the input image; obtaining a first output feature based on the encoding feature of the reference image and the high-frequency information; and obtaining a second output feature based on the encoding feature of the input image and the low-frequency information. The method further includes: generating an output image of a second resolution based on the first output feature and the second output feature, wherein the second resolution is greater than the first resolution.

Abstract: A method in an illustrative embodiment includes selecting, according to a type of input data, a target pre-trained model from a deep network pool including a plurality of pre-trained models; performing, by using the selected target pre-trained model, feature extraction on the input data to determine text descriptors for the input data; and generating, based on the text descriptors, a query table for query. The method can select, according to different input data, different target pre-trained models from the deep network pool including the plurality of pre-trained models to process (e.g., compress) the input data. The method assembles a plurality of deep networks into a pool to automatically process data to obtain text descriptors for data retrieval, thereby achieving efficient data compression and retrieval.

Abstract: Embodiments of the present disclosure provide a method for image generation for a particular view angle. The method comprises acquiring a three-dimensional scene model, a target camera pose, and a target view angle corresponding to a target scene. The method further comprises determining a target compressed image feature corresponding to the target camera pose and the target view angle from a plurality of compressed image features. The method further comprises inputting the target camera pose, the target view angle, and the target compressed image feature to the three-dimensional scene model, and obtaining a target image corresponding to the target camera pose and the target view angle through rendering by the three-dimensional scene model. By using embodiments of the present disclosure, it is possible to acquire a more accurate rendered image from a target view angle while saving the storage memory and increasing the loading speed.

Abstract: Example embodiments of the present disclosure provide a method, a device, and a computer program product for processing a workflow chart. The method includes encoding structural information of the workflow chart including a plurality of nodes and a plurality of edges by using a graph neural network to acquire a vector representation of the structural information; acquiring textual description data about the workflow chart at the nodes; training a language model based on the acquired textual description data and the acquired vector representation to acquire a pretrained language model; and fine-tuning the pretrained language model through training data of a specific task to acquire a fine-tuned language model. Through the method for processing the workflow chart of the present disclosure, the combination of the graph neural network and the language model not only can process a large number of complex workflow charts, but also can generate effective natural language outputs.

Abstract: Methods, electronic devices and computer program products for model processing are disclosed in embodiments herein. A method in an illustrative embodiment includes encoding first data of a point cloud model to obtain a first matrix, and encoding second data of the point cloud model to obtain a second matrix, where the first data and the second data are data of the point cloud model acquired from different angles. The method further includes respectively decomposing, by an equivariant encoder, the first matrix and the second matrix into a first equivariant matrix and a second equivariant matrix, and respectively decomposing, by an invariant encoder, the first matrix and the second matrix into a first invariant matrix and a second invariant matrix. The method further includes training the equivariant encoder and the invariant encoder based on the first equivariant matrix, the second equivariant matrix, the first invariant matrix, and the second invariant matrix.

Abstract: One example method includes obtaining a group of two dimensional images of a target, rendering, in real time immediately after the two dimensional images are obtained, a three dimensional model of the target, using the two dimensional images, and generating an output comprising the three dimensional model. The three dimensional model may then be compared with another version of the three dimensional model to identify a change in a condition of a structural element that is included in both of the three dimensional models.

Abstract: Techniques are disclosed for multi-factor prediction of computing resources for algorithm execution. For example, a method comprises obtaining a set of factors associated with an algorithm configured to transform one or more two-dimensional images into one or more three-dimensional models. The method further comprises computing an estimated computing power value based on the set of factors. The method then comprises scheduling execution of the algorithm on one or more computing resources based on the estimated computing power value.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for generating a three-dimensional image. The method includes: receiving a first image presenting a target object at a first viewing angle, wherein the first image is a two-dimensional image; determining a transformed image of the first image at a target viewing angle, wherein the target viewing angle is the same as or different from the first viewing angle. The method further includes: generating a first representation using a first feature extraction layer corresponding to the first viewing angle in an encoder based on the transformed image; and generating a second image based on the first representation, wherein the second image is a three-dimensional image and presents the target object at the target viewing angle.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for determining a node of a decision tree. The method includes determining multiple features of multiple modals corresponding to input information. The method further includes generating a multi-modal feature representation by combining the multiple features of the multiple modals. The method further includes determining a target path in a decision tree that is associated with the multi-modal feature representation, the decision tree comprising multiple nodes. The method further includes determining, in the target path based on the multi-modal feature representation, a target node associated with the input information and used to indicate a question or an answer. This method enables the fusion of feature representations corresponding to input information of different modals to determine a multi-modal feature representation. In this way, it is possible to determine richer and more accurate user intentions.

Abstract: Example embodiments of the present disclosure provide a method, a device, and a computer program product for data query. The method includes selecting, according to a type of input data, a target pre-trained model from a deep network pool including a plurality of pre-trained models; performing, by using the selected target pre-trained model, feature extraction on the input data to determine text descriptors for the input data; and generating, based on the text descriptors, a query table for query. The method according to the present disclosure can select, according to different input data, different target pre-trained models from the deep network pool including the plurality of pre-trained models to process (e.g., compress) the input data. The method according to the present disclosure assembles a plurality of deep networks into a pool to automatically process data to obtain text descriptors for data retrieval, thereby achieving efficient data compression and retrieval.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for video retrieval. The method includes determining a retrieval level corresponding to a retrieval word in response to receiving a retrieval request including the retrieval word from a client. The method further includes determining a video database corresponding to the retrieval level among a plurality of video databases, where the plurality of video databases store image frames for different frame rates of the same video. The method further includes retrieving an image frame associated with the retrieval word from the determined video database and sending the retrieved image frame to the client. According to the solution, multi-level video retrieval can be realized, allowing a user to retrieve a desired video in different scenes or different devices at different retrieval speeds, so as to provide the user with a more flexible video retrieval mode.