Abstract: An image processing method, a related device, and a computer storage medium are provided. The method includes: obtaining a feature intensity image corresponding to a training image, where an intensity value of a pixel in the feature intensity image is used to indicate importance of the pixel for recognizing the training image, and resolution of the training image is the same as resolution of the feature intensity image; and occluding, based on the feature intensity image, a to-be-occluded region in the training image by using a preset window, to obtain a new image, where the to-be-occluded region includes a to-be-occluded pixel, and the new image is used to update an image recognition model. According to the embodiments of the present application, a prior-art problem that a model has low accuracy and relatively poor generalization performance because of limited training data can be resolved.

Abstract: An image super-resolution method includes preprocessing the low-resolution image to obtain a vertical gradient map, a horizontal gradient map, and a luminance map, which are used as three different dimensions of information to constitute a to-be-input feature map, performing size conversion on the to-be-input feature map to obtain an input feature map, performing nonlinear transformation on the input feature map to obtain an input feature map obtained after the nonlinear transformation, and performing weighted processing on the input feature map and the input feature map obtained after the nonlinear transformation, to obtain an output feature map, performing size conversion on the output feature map to obtain a residual map, and combining the low-resolution image and the residual map to obtain a high-resolution image.

Abstract: A method for pedestrian attribute identification and positioning is provided. The method includes: performing feature extraction on a to-be-detected image at a plurality of different abstraction degrees, to obtain a plurality of first feature maps of a pedestrian attribute; performing convolution on the plurality of first feature maps, to obtain a plurality of second feature maps; mapping each second feature map to a plurality of areas (bins) that overlap each other, and performing max pooling on each bin, to obtain a plurality of high-dimensional feature vectors, where the plurality of bins that overlap each other evenly cover each second feature map; processing the plurality of high-dimensional feature vectors into a low-dimensional vector, to obtain an identification result of the pedestrian attribute; and further obtaining a positioning result of the pedestrian attribute based on the plurality of second feature maps and the plurality of high-dimensional feature vectors.

Abstract: A method and an apparatus for recognizing a descriptive attribute of an appearance feature include obtaining a location feature of an appearance feature of a target image to determine a location of a part of an object in a preset object model indicated by the appearance feature, where the location feature of the appearance feature indicates the location of the part of the object in the preset object model indicated by the appearance feature, recognizing a target region based on the location feature, where the target region includes the part of the object, performing feature analysis on the target region, recognizing a descriptive attribute of the appearance feature of the object, and determining the location feature of the appearance feature having a local attribute.

Abstract: Embodiments of the present invention disclose an image processing method, a related device, and a computer storage medium. The method includes: obtaining a feature intensity image corresponding to a training image, where an intensity value of a pixel in the feature intensity image is used to indicate importance of the pixel for recognizing the training image, and resolution of the training image is the same as resolution of the feature intensity image; and occluding, based on the feature intensity image, a to-be-occluded region in the training image by using a preset window, to obtain a new image, where the to-be-occluded region includes a to-be-occluded pixel, and the new image is used to update an image recognition model. According to the embodiments of the present invention, a prior-art problem that a model has low accuracy and relatively poor generalization performance because of limited training data can be resolved.

Abstract: An image super-resolution method includes preprocessing the low-resolution image to obtain a vertical gradient map, a horizontal gradient map, and a luminance map, which are used as three different dimensions of information to constitute a to-be-input feature map, performing size conversion on the to-be-input feature map to obtain an input feature map, performing nonlinear transformation on the input feature map to obtain an input feature map obtained after the nonlinear transformation, and performing weighted processing on the input feature map and the input feature map obtained after the nonlinear transformation, to obtain an output feature map, performing size conversion on the output feature map to obtain a residual map, and combining the low-resolution image and the residual map to obtain a high-resolution image.

Abstract: The method includes: performing feature extraction on a to-be-detected image at a plurality of different abstraction degrees, to obtain a plurality of first feature maps of a pedestrian attribute; performing convolution on the plurality of first feature maps, to obtain a plurality of second feature maps; mapping each second feature map to a plurality of areas (bins) that overlap each other, and performing max pooling on each bin, to obtain a plurality of high-dimensional feature vectors, where the plurality of bins that overlap each other evenly cover each second feature map; processing the plurality of high-dimensional feature vectors into a low-dimensional vector, to obtain an identification result of the pedestrian attribute; and further obtaining a positioning result of the pedestrian attribute based on the plurality of second feature maps and the plurality of high-dimensional feature vectors.

Abstract: A method and an apparatus for recognizing a descriptive attribute of an appearance feature include obtaining a location feature of an appearance feature of a target image to determine a location of a part of an object in a preset object model indicated by the appearance feature, where the location feature of the appearance feature indicates the location of the part of the object in the preset object model indicated by the appearance feature, recognizing a target region based on the location feature, where the target region includes the part of the object, performing feature analysis on the target region, recognizing a descriptive attribute of the appearance feature of the object, and determining the location feature of the appearance feature having a local attribute.

Abstract: A method, an apparatus, and a system for transmitting data are disclosed. In an embodiment, the includes receiving a detection interest packet, generating a detection response data packet according to a content name, and adding a mobility status of a first network node to a status record entry, sending the detection response data packet to a second network node so that the second network node adds a mobility status of the second network node to a first status record entry and sends the detection response data packet to a client. The client determines a revised value of a first expected distance according to the first status record entry and the first expected distance, generates a request interest packet, sends the request interest packet to at least one network node and receives the request interest packet sent by the second network node.

Abstract: Methods and apparatuses for routing and forwarding, establishing a routing table, and acquiring content are provided. The method for routing and forwarding includes: receiving a content request packet, where the content request packet carries a content name of requested content and container information of the requested content; determining whether any forwarding entry that matches the content name of the requested content exists in a FIB; determining whether any forwarding entry that matches the container identifier in the container information exists in the FIB when no forwarding entry that matches the content name of the requested content exists in the FIB; and when a forwarding entry that matches the container identifier in the container information of the requested content exists in the FIB, sending the content request packet according to a port in the matched forwarding entry.

Abstract: A method and an apparatus for compressing a content name are provided. In the method, a controller constructs a content name tree according to content names in a content database, where a first node of the content name tree stores a first-level prefix of a content name in the content database, and an Nth-level node of the content name tree stores prefixes of first N levels of the content name; and then the controller compresses a content name stored in the Mth-level node into a first content name stored in a father node of the Mth-level node when determining that the content name stored in each Mth-level node in the content name tree is corresponding to a first network node in a network. Thus, compression of a content name in a content name database is achieved, and space occupied by the content name is saved.

Abstract: A method, an apparatus, and a system for transmitting data are disclosed. In an embodiment, the includes receiving a detection interest packet, generating a detection response data packet according to a content name, and adding a mobility status of a first network node to a status record entry, sending the detection response data packet to a second network node so that the second network node adds a mobility status of the second network node to a first status record entry and sends the detection response data packet to a client. The client determines a revised value of a first expected distance according to the first status record entry and the first expected distance, generates a request interest packet, sends the request interest packet to at least one network node and receives the request interest packet sent by the second network node.

Abstract: A method is provided for content subscription in content network. A content requesting node in the content network receives a first interest packet comprising a content identifier through a first port from a first user device, creates a subscription entry comprising the content identifier and an identifier of the first port, sends the first interest packet to the content providing node, receives a second interest packet through a second port from a second user device; adds an identifier of the second port to the subscription entry, receives multiple data packets provided by the content providing node according to the first interest packet, and sends the multiple data packets through the first port and the second port respectively to the first user device and the second user device according to the subscription entry.

Abstract: A method and an apparatus for compressing a content name are provided. In the method, a controller constructs a content name tree according to content names in a content database, where a first node of the content name tree stores a first-level prefix of a content name in the content database, and an Nth-level node of the content name tree stores prefixes of first N levels of the content name; and then the controller compresses a content name stored in the Mth-level node into a first content name stored in a father node of the Mth-level node when determining that the content name stored in each Mth-level node in the content name tree is corresponding to a first network node in a network. Thus, compression of a content name in a content name database is achieved, and space occupied by the content name is saved.

Abstract: Embodiments of the present invention provide a container name server and a method for container name resolution. The container name server includes: a network interface configured to receive a resolution request packet sent by a network node, where the resolution request packet includes a container identification of a container to be resolved; and an execution engine configured to acquire a container identification of an access container of the container to be resolved according to the container identification of the container to be resolved. In embodiments of the present invention, possibilities for solving problems such as scalability and mobility support caused by routing based completely on a content name are provided by introducing a container name server and taking the container name server as a name resolution system in an ICN network.

Abstract: The present invention discloses a data source movement processing method, a packet forwarding method, and an apparatus thereof, which are applied to an information centric network (ICN). The data source movement processing method includes: when a target container from a first container enters a second container, registering a route of the target container in the second container; communicating with a resolution system, so that the resolution system updates an access container of the target container from the first container to the second container. In the data source movement processing method according to the embodiment of the present invention, a data source movement is supported without introducing frequent and cumbersome route updates and without changing a content name of content, thereby maintaining persistence of the content name, enhancing feasibility of an ICN architecture, and improving user experience.

Abstract: Methods and apparatuses for routing and forwarding, establishing a routing table, and acquiring content are provided. The method for routing and forwarding includes: receiving a content request packet, where the content request packet carries a content name of requested content and container information of the requested content; determining whether any forwarding entry that matches the content name of the requested content exists in a FIB; determining whether any forwarding entry that matches the container identifier in the container information exists in the FIB when no forwarding entry that matches the content name of the requested content exists in the FIB; and when a forwarding entry that matches the container identifier in the container information of the requested content exists in the FIB, sending the content request packet according to a port in the matched forwarding entry.

Abstract: A method for controlling uplink power of an LTE system, a base station, and a mobile terminal are disclosed according to the embodiments of the present invention. The method includes: obtaining uplink transmit power of each carrier; sending first notification signaling to a base station if a difference between the obtained uplink transmit power of each carrier exceeds a specific threshold; and sending second notification signaling to the base station if a sum of the obtained uplink transmit power of each carrier exceeds a preset threshold. Thresholds are determined according to network conditions, and uplink power of a carrier is adjusted in a balanced way, a bit error rate is reduced and a case that a low-performance carrier is forced to be interrupted is prevented.

Abstract: A method is provided for content subscription in content network. As disclosed in the method, a content requesting node in the content network receives a first interest packet comprising a content identifier through a first port from a first user device, creates a subscription entry comprising the content identifier and an identifier of the first port, sends the first interest packet to the content providing node, receives a second interest packet through a second port from a second user device; adds an identifier of the second port to the subscription entry, receives multiple data packets provided by the content providing node according to the first interest packet, and sends the multiple data packets through the first port and the second port respectively to the first user device and the second user device according to the subscription entry.

Abstract: Embodiments of the present invention disclose a downlink Inter-Cell Interference Coordination (ICIC) method and relate to the wireless communication field. The method includes: obtaining information about multiple carriers in a current cell; obtaining, according to the information about the multiple carriers, a downlink ICIC parameter corresponding to each carrier of the multiple carriers; and sending the downlink ICIC parameter corresponding to each carrier to a neighboring cell of the current cell so as to enable the neighboring cell to perform, according to the downlink ICIC parameter, downlink interference coordination with the first cell. The embodiments of the present invention are applicable to downlink ICIC in a multi-carrier system.