Abstract: The present invention relates to a full-electronic turnout security control system for a train-to-train communication and a method thereof. The system comprises: a communication module for realizing an external communication respectively with a resource manager and a maintenance system and an internal communication respectively with a turnout control module, a relay driving module and a relay collection module; the turnout control module for controlling rotation of a turnout and collecting a turnout representation; the relay driving module for driving a relay to turn on or off according to a turnout power source on or off command transmitted by the resource manager; the relay collection module for collecting an on or off state of the collection relay and respectively transmitting the state to the resource manager and the maintenance system through the communication module; and a relay for connecting the relay driving module, the relay collection module and a switch machine driving return line.

Abstract: The disclosure relates to a non-national standard turnout drive system based on a double 2-vote-2 architecture, including an interlocking processing subsystem IPS, an interlocking maintenance station SDM, a non-national standard turnout drive module HIOM and an interlocking maintenance station SDM, wherein the non-national standard turnout drive module HIOM, a full-electronic communication module EIOCOM2, and the interlocking processing subsystem IPS are connected with each other in order, and the full-electronic communication module EIOCOM2 is connected to the interlocking maintenance station SDM; two non-national standard turnout drive module HIOMs, which are mutually redundant, obtain turnout drive commands through the interlocking processing subsystem IPS to control drive relays in a non-national standard turnout to lift and fall for driving the turnout to rotate toward a specified direction, while collecting representation information of the turnout and determining a position of the turnout.

Abstract: A method for generating semantic information may include obtain a scanning image. The scanning image may include a plurality of pixels representing an anatomical structure. The method may also include obtain a trained segmentation model. The method may further include determine a location probability distribution of the anatomical structure in the scanning image based on the trained segmentation model. The method may also include generate a segmentation result related to the anatomical structure based on the location probability distribution. The method may further include save the segmentation result into a tag of a digital imaging and communications in medicine (DICOM) file.

Abstract: The disclosure relates to a non-national standard turnout drive system based on a double 2-vote-2 architecture, including an interlocking processing subsystem IPS, an interlocking maintenance station SDM, a non-national standard turnout drive module HIOM and an interlocking maintenance station SDM, wherein the non-national standard turnout drive module HIOM, a full-electronic communication module EIOCOM2, and the interlocking processing subsystem IPS are connected with each other in order, and the full-electronic communication module EIOCOM2 is connected to the interlocking maintenance station SDM; two non-national standard turnout drive module HIOMs, which are mutually redundant, obtain turnout drive commands through the interlocking processing subsystem IPS to control drive relays in a non-national standard turnout to lift and fall for driving the turnout to rotate toward a specified direction, while collecting representation information of the turnout and determining a position of the turnout.

Abstract: A method for generating semantic information may include obtain a scanning image. The scanning image may include a plurality of pixels representing an anatomical structure. The method may also include obtain a trained segmentation model. The method may further include determine a location probability distribution of the anatomical structure in the scanning image based on the trained segmentation model. The method may also include generate a segmentation result related to the anatomical structure based on the location probability distribution. The method may further include save the segmentation result into a tag of a digital imaging and communications in medicine (DICOM) file.