Abstract: Disclosed in the embodiments of the present disclosure are a desensitization method and apparatus for image data, wherein the method comprises: determining a target image channel, which needs to be subjected to target recognition, in a native image; performing image restoration on native data corresponding to the target image channel, so as to obtain a single-channel image corresponding to a color component of the target image channel; recognizing a target object from the single-channel image; determining a sensitive area in the native image on the basis of the target object; performing data desensitization processing on the native data in the sensitive area. By the embodiments of the present disclosure, it can be guaranteed that privacy image data cannot be stored and leaked in a subsequent digital image signal processing process, thereby providing protection for the privacy data from the start of the generation of image data and improving data security.

Abstract: A method comprises allocating a data buffer in a shared memory region of a computing device, writing communication data comprising a header portion and a body portion into the data buffer, generating a modified header portion by appending the header portion of the communication data with address data that determines an address of the communication data in the data buffer, processing the modified header portion through the network protocol stack, and storing the body portion in the data buffer in a compatible data format for communication according to the networking protocol stack.

Abstract: A method comprises allocating a data buffer in a shared memory region of a computing device, writing communication data comprising a header portion and a body portion into the data buffer, generating a modified header portion by appending the header portion of the communication data with address data that determines an address of the communication data in the data buffer, processing the modified header portion through the network protocol stack, and storing the body portion in the data buffer in a compatible data format for communication according to the networking protocol stack.

Abstract: Disclosed are a method for updating a neural network and an electronic device. The method includes: inputting a first image set having tag information into a first depth neural network, and determining a cross entropy loss value of the first image set by using the first depth neural network; inputting a second image set having no tag information separately into the first depth neural network and a second depth neural network, and determining a consistency loss value of the second image set, the first depth neural network and the second depth neural network having the same network structure; updating parameters of the first depth neural network based on the cross entropy loss value and the consistency loss value; and updating parameters of the second depth neural network based on the updated parameters of the first depth neural network.

Abstract: A data processing method includes determining whether or not it is possible to perform a new processing task within the new processing task's time limit if execution of one or more current processing tasks is continued based on the one or more current processing tasks' configurations. In case the new task's performing is determined not to be possible within the new task's time limit, resources are released from the current tasks so as to still enable performing the current tasks within their respective time limits.

Abstract: An image recognition method includes: determining a first feature map of the current frame image by using a convolutional neural network based on a type of a current frame image; determining a second feature map of a key frame image before the current frame image; performing feature alignment on the first feature map and the second feature map to obtain a first aligned feature map; fusing the first feature map and the first aligned feature map to obtain a first fused feature map; and recognizing content in the current frame image based on the first fused feature map.

Abstract: A data processing method includes determining whether or not it is possible to perform a new processing task within the new processing task's time limit if execution of one or more current processing tasks is continued based on the one or more current processing tasks' configurations. In case the new task's performing is determined not to be possible within the new task's time limit, resources are released from the current tasks so as to still enable performing the current tasks within their respective time limits.

Abstract: A computer architecture for labeling documents is disclosed. According to some aspects, a computer accesses a collection of documents corresponding to a medical encounter and a labeling for the collection, wherein the labeling comprises one or more labels representing medical annotations assigned to the medical encounter. The computer computes, using a Hierarchical Attention Network (HAN), for each of a plurality of document-label pairs, a probability that a document of the document-label pair corresponds to a label of the document-label pair based on one or more features of text in the document, wherein each document-label pair comprises a document from the collection of documents and a label from the labeling. The computer provides an output representing the computed probabilities.

Abstract: An image recognition method includes: determining a first feature map of the current frame image by using a convolutional neural network based on a type of a current frame image; determining a second feature map of a key frame image before the current frame image; performing feature alignment on the first feature map and the second feature map to obtain a first aligned feature map; fusing the first feature map and the first aligned feature map to obtain a first fused feature map; and recognizing content in the current frame image based on the first fused feature map.

Abstract: Disclosed are a method for updating a neural network and an electronic device. The method includes: inputting a first image set having tag information into a first depth neural network, and determining a cross entropy loss value of the first image set by using the first depth neural network; inputting a second image set having no tag information separately into the first depth neural network and a second depth neural network, and determining a consistency loss value of the second image set, the first depth neural network and the second depth neural network having the same network structure; updating parameters of the first depth neural network based on the cross entropy loss value and the consistency loss value; and updating parameters of the second depth neural network based on the updated parameters of the first depth neural network.

Abstract: This disclosure describes a computerized system for managing medical information, the system including at least one computing device configured to receive a plurality of medical documents, determine skip grams comprising tokens for the plurality of medical documents, wherein each of the skip grams comprises one or more tokens of at least one of the medical documents, and populate a trie data structure based on the skip grams. The at least one computing device is further configured to prune the plurality of nodes of the trie based on one or more criteria to produce a pruned trie of nodes, determine rules for associating medical billing codes with the skip grams of the pruned trie of nodes based on pointwise mutual information, and output the determined rules.

Abstract: Embodiments of the disclosure include methods of prenatal testing using non-invasive means that identify single gene disorders. In specific embodiments the methods are non-invasive and employ tagging circulating cell-free fetal DNA from the biological mother with particular adaptors that employ unique barcodes, followed by steps to enrich targets and steps for thorough sequencing.

Abstract: This disclosure describes a computerized system for managing medical information, the system including at least one computing device configured to receive a plurality of medical documents, determine skip grams comprising tokens for the plurality of medical documents, wherein each of the skip grams comprises one or more tokens of at least one of the medical documents, and populate a trie data structure based on the skip grams. The at least one computing device is further configured to prune the plurality of nodes of the trie based on one or more criteria to produce a pruned trie of nodes, determine rules for associating medical billing codes with the skip grams of the pruned trie of nodes based on pointwise mutual information, and output the determined rules.

Abstract: An Organic Light Emitting Diode (OLED) display panel is disclosed. The display panel includes a substrate, a plurality of power lines disposed on the substrate, and a reflection layer disposed on the power lines, where the reflection layer is electrically connected with the power lines. The display panel also includes an anode disposed on the reflection layer, an optical modulation layer disposed between the reflection layer and the anode, a cathode disposed on the anode, and an organic emitting device layer disposed between the anode and the cathode, where the reflection layer is insulated from the anode, and the OLED display panel is configured to transmit light from a side of the reflection layer away from the substrate.

Abstract: An Organic Light Emitting Diode (OLED) display panel is disclosed. The display panel includes a substrate, a plurality of power lines disposed on the substrate, and a reflection layer disposed on the power lines, where the reflection layer is electrically connected with the power lines. The display panel also includes an anode disposed on the reflection layer, an optical modulation layer disposed between the reflection layer and the anode, a cathode disposed on the anode, and an organic emitting device layer disposed between the anode and the cathode, where the reflection layer is insulated from the anode, and the OLED display panel is configured to transmit light from a side of the reflection layer away from the substrate.

Abstract: A five volt tolerant integrated circuit signal pad having an initial fast pull-up to three volts and then operating as an open drain output with an external resistor for pulling up the output from about three volts to about five volts. The initial fast (active) pull-up is accomplished with active devices that reduces the overall pull-up time of a newer technology (lower operating voltage) integrated circuit output when transitioning from a logic 0 to a logic 1. Circuits of the integrated circuit output driver protect the internal operating circuit nodes from excessive voltage and leakage currents that would otherwise result from a voltage on the signal pad that is more positive than the operating voltage of the integrated circuit.