A method disclosed herein provides for receiving, at a user device, a media stream including frames of a first resolution generated by a graphics-rendering application and utilizing one or more weight matrices pre-trained in association with the graphics-rendering application to locally upscale each received frame of the media stream at the user device to a second resolution greater than the first resolution. Local upscaling of the media stream may be performed “on the fly,” such as with respect to individual content streams (e.g., a game) or segments of content streams (e.g., a scene within a game).

The present invention relates to a method for playing on a player of a client device (11) a content streamed in a network (1), said content consisting of a sequence of segments available in a plurality of quality levels, the player being configured so as to choose the quality level of the segments as a function of at least one parameter representative of a segment reception rate, according to an Adaptive BitRate, ABR, logic of the player; the client device (11) comprising a first buffer (M1) for storing segments in a format adapted for transferring within the network (1), the method being characterized in that it comprises performing by a processing unit (110) of the client device (11): (a) receiving from the player a request for a current segment at a first quality level; (b) determining that the player will request according to its ABR logic a next segment at a second quality level after said requested current segment is provided from the first buffer memory (M1), using a model predicting the ABR logic of the player; (c) if said next segment is not present at said second quality level in the first buffer memory (M1), fetching said next segment at said second quality level from the network (1).

Methods, apparatus, systems, and articles of manufacture are disclosed for the detection of boring media. An example apparatus includes at least one memory, instructions, and processor circuitry to correlate first boring media event determinations with a first portion of first media events, train a machine learning model based on the first boring media event determinations, the machine learning model to predict second boring media event determinations associated with a second portion of the first media events, and, in response to a number of the second boring media event determinations correctly predicted by the machine learning model satisfying a threshold, deploy the machine learning model to predict third boring media event determinations associated with second media events.

A method of detecting an action, an electronic device, and a storage medium. A method can include: performing a temporal action proposal on at least one target feature data obtained by a feature extraction on a plurality of target frame data of a target resource, so as to obtain at least one first candidate action proposal information; classifying target feature data corresponding to at least one first candidate action proposal interval included in the first candidate action proposal information, so as to obtain at least one classification confidence level corresponding to the at least one first candidate action proposal interval; and determining an action detection result for at least one action segment contained in the target resource according to the at least one classification confidence level corresponding to the at least one first candidate action proposal interval, wherein the action detection result includes an action category and an action period.

Systems and methods for generating individualized content trailers. Content such as a video is divided into segments each representing a set of common features. With reference to a set of stored user preferences, certain segments are selected as aligning with the user's interests. Each selected segment may then be assigned a label corresponding to the plot portion or element to which it belongs. A coherent trailer may then be assembled from the selected segments, ordered according to their plot elements. This allows a user to see not only segments containing subject matter that aligns with their interests, but also a set of such segments arranged to give the user an idea of the plot, and a sense of drama, increasing the likelihood of engagement with the content.

Provided herein are methods and systems for improved accessibility scoring for content items. A predicted accessibility score may be based on a plurality of multimodal features present within a content item. The plurality of multimodal features may include video features (e.g., based on video/image analysis), audio features (e.g., based on audio analysis), text-based features (e.g., based on closed-captioning analysis), features indicated by metadata (e.g., duration, genre, etc.), a combination thereof, and/or the like. A predicted accessibility score for a content item may indicate how accessible the content item may be for persons who are visually impaired, hearing impaired, cognitively impaired, etc., as well as for persons who desire to view content that requires less visual attention and/or audio attention as the case may be.

A method of operating a display apparatus includes: obtaining situation information for voice recognizer selection, selecting at least one of a plurality of voice recognizers based on the situation information, obtaining a voice recognition result from a voice signal, using the selected at least one voice recognizer, and obtaining a chat message from the voice recognition result.

A method and an electronic device for detecting the resolution of a video material are provided. The method is applied to an electronic device, and a calculation circuit of the electronic device executes an AI model. The video material includes multiple frames, and each frame includes multiple sub-frames. The AI model processes multiple pixel data to generate an intermediate resolution corresponding to the pixel data. The method includes the following steps: (A) generating a target sub-frame, the number of pixels of the target sub-frame being smaller than the number of pixels in any frame; (B) inputting the target sub-frame into the AI model to generate the intermediate resolution; (C) storing the intermediate resolution; (D) repeating steps (A) to (C) to generate multiple intermediate resolutions; and (E) determining the resolution of the video material based on the intermediate resolutions.

Systems and methods are described for selecting content item identifiers for display. The system may identify a set of content items that are likely to be requested in the future based on a history of content item requests. The system then selects a first plurality of content categories using a category selection neural net and selects a first set of recommended content items for the first plurality of content categories. The system increases a reward score for the first plurality of content categories based on receiving a request for a content item that is included in the first set of recommended content items. The system also decreases the reward score for the first plurality of content categories based on determining that the requested content item is included in the set of content items that are likely to be requested in the future. The neural net is trained based on the reward score of the first plurality of content categories to reinforce reward score maximization. The trained neural net is the used to select content items for display.

An electronic device according to an embodiment of the disclosure includes: a communicator; a memory storing one or more instructions; at least one processor configured to execute the one or more instructions stored in the memory to collect content metadata and user metadata from a plurality of different servers that provide content, obtain a content latent factor including information about similarities between pieces of the content based on characteristics of the content metadata, by using a first learning network model, obtain a user latent factor related to user preferred content information based on characteristics of the user metadata, by using a second learning network model, obtain a user preference score for the content based on the content latent factor and the user latent factor, by using a third learning network model, and provide a recommended content list based on the user preference score.