Abstract: Provided are a facial expression recognition method and system combined with an attention mechanism. The method comprises: detecting faces comprised in each video frame in a video sequence, and extracting corresponding facial ROIs, so as to obtain facial pictures in each video frame; aligning the facial pictures in each video frame on the basis of location information of facial feature points of the facial pictures; inputting the aligned facial pictures into a residual neural network, and extracting spatial features of facial expressions corresponding to the facial pictures; inputting the spatial features of the facial expressions into a hybrid attention module to acquire fused features of the facial expressions; inputting the fused features of the facial expressions into a gated recurrent unit, and extracting temporal features of the facial expressions; and inputting the temporal features of the facial expressions into a fully connected layer, and classifying and recognizing the facial expressions.

Abstract: Provided is a method for preparing Rebaudioside J using an enzymatic method, comprising using rebaudioside A as a substrate, and making the substrate, in the presence of a glycosyl donor, react under the catalysis of a UDP-glycosyltransferase-containing recombinant cell and/or UDP-glycosyltransferase prepared therefrom to generate Rebaudioside J.

Abstract: A light-emitting device includes a semiconductor epitaxial structure including a first semiconductor layer, an active layer, and a second semiconductor layer, and having holes; a first insulation layer disposed on the semiconductor epitaxial structure and having first and second grooves; a first pad electrically connected to the first semiconductor layer through the first grooves; and a second pad electrically connected to the second semiconductor layer through the second grooves. A projection of the first pad does not overlap projections of the holes. A projection of the second pad does not overlap the projections of the holes. The first pad includes a first pad connection portion and first pad extension portions; the second pad includes a second pad connection portion and second pad extension portions. Projections of the second grooves fall between projections of the first and second pad extension portions. Two other aspects of the light-emitting device are also provided.

Abstract: The present disclosure provides a non-contact fatigue detection system and method based on rPPG. The system and method adopt multi-thread synchronous communication for real-time acquisition and processing of rPPG signal, enabling fatigue status detection. In this setup, the first thread handles real-time rPPG data capture, storage and concatenation, while the second thread conducts real-time analysis and fatigue detection of rPPG data. Through a combination of skin detection and LUV color space conversion, rPPG raw signal extraction is achieved, effectively eliminating interference from internal and external environmental facial noise; Subsequently, an adaptive multi-stage filtering process enhances the signal-to-noise ratio, and a multi-dimensional fusion CNN model ensures accurate detection of respiration and heart rate.

Abstract: Disclosed are a non-contact fatigue detection method and system. The method comprises: sending a millimeter-wave (mmWave) radar signal to a person being detected, receiving an echo signal reflected from the person, and determining a time-frequency domain feature, a non-linear feature and a time-series feature of a vital sign signal; acquiring a facial video image of the person, and performing facial detection and alignment on the basis of the facial video image, for extracting a time domain feature and a spatial domain feature of the person's face; fusing the determined vital sign signal with the time domain feature and the spatial domain feature of the person's face, for obtaining a fused feature; and determining whether the person is in a fatigued state by the fused feature. By fusing the two detection techniques, the method effectively suppressing the interference of subjective and objective factors, and improving the accuracy of fatigue detection.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for backup. The method includes: determining a plurality of buffer periods associated with a plurality of pending backup operations of a plurality of objects, each buffer period being a preprocessing period for a corresponding pending backup operation. The method further includes: determining a plurality of predicted execution durations of the plurality of pending backup operations based on historical execution durations of respective historical backup operations of the plurality of objects. The method further includes: determining priorities of the plurality of pending backup operations based on the plurality of predicted execution durations and the plurality of buffer periods. The method further includes: executing the plurality of pending backup operations based on the priorities.

Abstract: Provided are a facial expression recognition method and system combined with an attention mechanism. The method comprises: detecting faces comprised in each video frame in a video sequence, and extracting corresponding facial ROIs, so as to obtain facial pictures in each video frame; aligning the facial pictures in each video frame on the basis of location information of facial feature points of the facial pictures; inputting the aligned facial pictures into a residual neural network, and extracting spatial features of facial expressions corresponding to the facial pictures; inputting the spatial features of the facial expressions into a hybrid attention module to acquire fused features of the facial expressions; inputting the fused features of the facial expressions into a gated recurrent unit, and extracting temporal features of the facial expressions; and inputting the temporal features of the facial expressions into a fully connected layer, and classifying and recognizing the facial expressions.

Abstract: A current retention mode is identified as being one of conservative or aggressive. When the retention mode is identified as aggressive, a check is made as to whether used space on backup storage is below a lower threshold. If the used space is below the lower threshold, a switch is made into the conservative mode. If the used space is above the lower threshold, the aggressive mode is maintained. While in the aggressive mode, a backup chain of a database is selected for pruning, the selected backup chain having a shortest retention term of all backup chains not yet pruned and not being a latest backup chain. Log backups of the database in the selected backup chain are identified. Each log backup that has expired is deleted regardless of dependency of other backups on an expired log backup.

Abstract: A method, electronic device, and computer-readable medium for storage management is disclosed. The method for storage management includes acquiring a data change rate of a data source, the data change rate indicating an occurrence rate of data to be backed up in the data source. The method also includes acquiring a recovery capability of a backup system to recover backed-up data and determining, based on the data change rate and the recovery capability, a backup strategy for backing up the data to be backed up.

Abstract: A system comprising at least one hardware processor for updating a cloud-based road anomaly database is disclosed, wherein the system receives information from a vehicle regarding an object detected by one or more sensors of the vehicle, the information may include a location and/or a size of the detected object; compares the location and/or the size of the first detected object against data stored in a road feature database regarding the first detected object; determines, based on comparing the location and/or the size, an accuracy score associated with the first detected object and the second detected object; and updates the cloud-based road anomaly database with the received information for the second detected object based on the associated accuracy score being higher than at least one of a threshold accuracy score or an accuracy score associated with corresponding object information stored in the cloud-based road anomaly database.

Abstract: In some examples, a system may determine a plurality of candidate routes between a source location of a vehicle and a destination location for the vehicle. Further, the system may determine one or more respective observation zone volumes for each candidate route of the plurality of candidate routes. The system may determine a field of view (FOV) for one or more vehicle sensors onboard the vehicle. In addition, the system may select a route for the vehicle from the plurality of candidate routes based at least on comparing an overlap of the FOV with the respective observation zone volumes for the candidate routes.

Abstract: A current retention mode is identified as being one of conservative or aggressive. When the retention mode is identified as aggressive, a check is made as to whether used space on backup storage is below a lower threshold. If the used space is below the lower threshold, a switch is made into the conservative mode. If the used space is above the lower threshold, the aggressive mode is maintained. While in the aggressive mode, a backup chain of a database is selected for pruning, the selected backup chain having a shortest retention term of all backup chains not yet pruned and not being a latest backup chain. Log backups of the database in the selected backup chain are identified. Each log backup that has expired is deleted regardless of dependency of other backups on an expired log backup.

Abstract: A method, electronic device, and computer-readable medium for storage management is disclosed. The method for storage management includes acquiring a data change rate of a data source, the data change rate indicating an occurrence rate of data to be backed up in the data source. The method also includes acquiring a recovery capability of a backup system to recover backed-up data and determining, based on the data change rate and the recovery capability, a backup strategy for backing up the data to be backed up.

Abstract: Embodiments of the present disclosure relate to a method, device and computer program product for determining a consistency level of a virtual machine snapshot. The method comprises obtaining a backup log file of the virtual machine snapshot generated in a virtual machine backup process, the backup log file comprising log information related to the generation of the virtual machine snapshot. The method further comprises determining the consistency level of the virtual machine snapshot based on the backup log file, the consistency level indicating a matching degree between a state of a virtual machine represented by the virtual machine snapshot and an actual state of the virtual machine when the virtual machine snapshot is generated.

Abstract: Embodiments of the present disclosure relate to a method for managing backup jobs, an electronic device, and a computer program product. The method includes: determining expected execution durations of a group of to-be-executed backup jobs; dividing the group of to-be-executed backup jobs into a plurality of backup job subsets based on the expected execution durations, wherein a difference between the expected execution durations of every two backup jobs in each backup job subset does not exceed a predetermined threshold duration; and adjusting an execution plan of the group of to-be-executed backup jobs to cause the backup jobs in at least one backup job subset in the plurality of backup job subsets to simultaneously begin to be executed.

Abstract: An efficient method for optimizing long-term retention backup policy within recovery point objectives (RPO). Specifically, the disclosed method proposes a dynamic promotion scheme through which short-term retention backup copies, in compliance with specified long-term retention RPOs, may be promoted to render long-term retention backup copies. Further, the disclosed method not only looks to past and/or presently dated short-term retention backup copies, but also looks to prospective (or future) dated short-term retention backup copies, which are expected or predicted to be produced, for promotion. Moreover, in circumstances where there are no appropriate past, present, or future dated short-term retention backup copies to promote, the disclosed method triggers new backup operations to acquire the long-term retention backup copies necessary to maintain the specified long-retention RPOs.

Abstract: A efficient method for long-term retention backup policy within recovery point objectives (RPO). Specifically, the disclosed method proposes a dynamic promotion scheme through which short-term retention backup copies, in compliance with specified long-term retention RPOs, may be promoted to render long-term retention backup copies. Further, the disclosed method not only looks to past and/or presently dated short-term retention backup copies, but also looks to prospective (or future) dated short-term retention backup copies, which are expected or predicted to be produced, for promotion. Moreover, in circumstances where there are no appropriate past, present, or future dated short-term retention backup copies to promote, the disclosed method triggers new backup operations to acquire the long-term retention backup copies necessary to maintain the specified long-retention RPOs.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for determining an execution progress of tasks. The method includes determining, according to a determination that a task is executed, whether historical execution information of the task is available. The method further includes determining the expected execution duration of the task based on the historical execution information of the task according to a determination that the historical execution information of the task is available. The method further includes determining the duration of completed execution for the task based on the time point at which execution of the task begins and the current time point. The method further includes determining the execution progress of the task based on the expected execution duration and the completed execution duration.

Abstract: A method and system for performing authentication for a backup service provided by a server is provided. The method receives a request for authentication from a client device, the request for authentication including a signature generated using a private key. The method sends a request to obtain a public key corresponding to the private key to the server and receives the public key from the server, the public key being retrieved by the server from a backup of a virtual machine. The method verifies the signature using the public key and generates a token encrypted using the public key, the token enabling the client device to access the server for the backup service. The method sends the token to the client device, the token to be decrypted using the private key by the client device.

Abstract: Embodiments of the present disclosure provide a data backup method, an electronic device and a computer program product. The method provided includes creating a workflow related to a data source, the workflow being configured to generate a copy of the data source or to replicate the copy of the data source to a remote device. The method further includes identifying a dependency of the workflow, the dependency indicating whether the workflow depends on at least one additional workflow related to the data source. The method further includes scheduling the workflow based at least on the identified dependency. According to the embodiments of the present disclosure, by use of a dependency-based scheduling mechanism, the performance of a storage system can be improved, and consumption of resources such as a storage and an internal memory can be reduced.