Abstract: The present disclosure provides an image processing method and apparatus, and an electronic device, and relates to the field of artificial intelligence (AI). The method includes: obtaining an input image and extracting region features of image regions in the input image to obtain a first image feature; processing pixels in the first image feature according to a predetermined rule, and determining a second image feature according to the processed pixels; and determining, based on the second image feature and at least one word vector that is determined for the input image, word vectors corresponding to the region features of the image regions in the first image feature at different moments, predicting positions of the word vectors in a text description, and forming the text description corresponding to the input image according to the word vectors and the positions.

Abstract: A method for developing an epoxy resin impregnated glass fiber Direct Current (DC) bushing, comprising: according to length parameters of each layer of capacitive screen or resistive screen designed depending on insulation requirements, selecting bushing design parameters, determining a winding machine program according to the bushing design parameters, and winding a core body according to the winding machine program, wherein during the core body winding process, the core body begins to be initially cured; after the core body is wound, curing the core body by an oven according to a preset oven temperature and duration; machining the cured core body according to a preset core body design drawing; after the inner wall of a flange is polished and cleaned and is heated and pretreated by the oven, injecting glue at the position of a glue injection hole of the flange for gluing the core body and the flange; sequentially assembling a collector ring, a hollow composite insulator, and a voltage-equalizing sealing cove

Abstract: A video caption generating method is provided to a computer device. The method includes encoding a target video by using an encoder of a video caption generating model, to obtain a target visual feature of the target video, decoding the target visual feature by using a basic decoder of the video caption generating model, to obtain a first selection probability corresponding to a candidate word, decoding the target visual feature by using an auxiliary decoder of the video caption generating model, to obtain a second selection probability corresponding to the candidate word, a memory structure of the auxiliary decoder including reference visual context information corresponding to the candidate word, determining a decoded word in the candidate word according to the first selection probability and the second selection probability, and generating a video caption according to decoded word.

Abstract: The present disclosure provides an image processing method and apparatus, and an electronic device, and relates to the field of artificial intelligence (AI). The method includes: obtaining an input image and extracting region features of image regions in the input image to obtain a first image feature; processing pixels in the first image feature according to a predetermined rule, and determining a second image feature according to the processed pixels; and determining, based on the second image feature and at least one word vector that is determined for the input image, word vectors corresponding to the region features of the image regions in the first image feature at different moments, predicting positions of the word vectors in a text description, and forming the text description corresponding to the input image according to the word vectors and the positions.

Abstract: A video caption generating method is provided to a computer device. The method includes encoding a target video by using an encoder of a video caption generating model, to obtain a target visual feature of the target video, decoding the target visual feature by using a basic decoder of the video caption generating model, to obtain a first selection probability corresponding to a candidate word, decoding the target visual feature by using an auxiliary decoder of the video caption generating model, to obtain a second selection probability corresponding to the candidate word, a memory structure of the auxiliary decoder including reference visual context information corresponding to the candidate word, determining a decoded word in the candidate word according to the first selection probability and the second selection probability, and generating a video caption according to decoded word.

Abstract: Certain aspects of the present disclosure relate to methods and apparatus for management of hybrid automatic repeat request (HARQ) log likelihood ratio (LLR) and reordering buffers in wireless communication systems. According to certain aspects, a method for reducing buffer overhead that may be performed by a wireless node is provided. The method generally includes receiving one or more packets of at least one of an initial transmission or a retransmission; forming one or more log likelihood ratios (LLRs) based on the one or more packets; compressing the one or more LLRs by quantizing the one or more LLRs; and buffering the one or more compressed LLRs.

Abstract: Certain aspects of the present disclosure relate to methods and apparatus for management of hybrid automatic repeat request (HARQ) log likelihood ratio (LLR) and reordering buffers in wireless communication systems. According to certain aspects, a method for reducing buffer overhead that may be performed by a wireless node is provided. The method generally includes receiving one or more packets of at least one of an initial transmission or a retransmission; forming one or more log likelihood ratios (LLRs) based on the one or more packets; compressing the one or more LLRs by quantizing the one or more LLRs; and buffering the one or more compressed LLRs.

Abstract: Certain aspects relate to methods and apparatus for obtaining system information by an apparatus comprising concurrently maintaining a first buffer for combining multiple transmissions of at least a first type of system information block (SIB) message across different system information (SI) message windows and a second buffer for combining multiple transmissions of at least a second type of SIB message within an SI window, and decoding at least first and second types of SIB messages based on contents in the first and second buffers.

Abstract: A method, an apparatus, and a computer program product for wireless communication are provided. The apparatus may be a UE. The UE may search for one or more cells during each of a number of search periods, select a first cell that has been detected in at least two of the search periods, and determine an LNA gain based on information associated with the first cell.

Abstract: Pipeline automatic gain control (AGC) processing is conducted by a mobile device when preparing for reception of one or more synchronization signals in a cell, when the mobile device has no knowledge of a timing and/or configuration of the synchronization signals in the cell. The mobile device triggers selection of a particular set of pipeline AGC processing operations and sets an amplifier with an amplifier gain setting determined by the selected set of pipeline AGC processing operations.