Abstract: The present disclosure relates to a network optimization method and apparatus, an image processing method and apparatus, and a storage medium. The network optimization method includes: obtaining an image sample group; obtaining a first feature and a second feature of an image in the image sample group, and obtaining a first classification result by using the first feature of the image; performing feature exchange processing on an image pair in the image sample group to obtain a new image pair; obtaining a first loss value of the first classification result, a second loss value of the new image pair, and a third loss value of first features and second features of the new image pair in a preset manner; and adjusting parameters of a neural network at least according to the first loss value, the second loss value, and the third loss value until a preset requirement is met.

Abstract: The present disclosure discloses a rotary joint electromagnetic locking device and a rotary joint, the rotary joint electromagnetic locking device comprising a first electromagnetic sucker, a sliding plate, a rotary plate, a connecting shaft connecting the first electromagnetic sucker and the sliding plate, and a second electromagnetic sucker, different from the prior art. When the first electromagnetic sucker and the second electromagnetic sucker have no adsorption force, the sliding plate and the rotary plate can be locked to each other; when the second electromagnetic sucker has an adsorption force, the rotary plate or the first electromagnetic sucker is adsorbed by the second electromagnetic sucker, making the sliding plate and the rotary plate to be further locked; when the first electromagnetic sucker has an adsorption force, the sliding plate can be adsorbed to the first contact surface of the first electromagnetic sucker, making the sliding plate released from the rotary plate.

Abstract: A method for identifying a target, a non-transitory computer-readable storage medium, and an electronic device include: acquiring a first image and a second image, the first image and the second image each including a target to be determined; generating a prediction path based on the first image and the second image, both ends of the prediction path respectively corresponding to the first image and the second image; and performing validity determination on the prediction path and determining, according to a determination result, whether the targets to be determined in the first image and the second image are the same target to be determined.

Abstract: A method for identifying a target, a non-transitory computer-readable storage medium, and an electronic device include: acquiring a first image and a second image, the first image and the second image each including a target to be determined; generating a prediction path based on the first image and the second image, both ends of the prediction path respectively corresponding to the first image and the second image; and performing validity determination on the prediction path and determining, according to a determination result, whether the targets to be determined in the first image and the second image are the same target to be determined.

Abstract: Techniques for building a backward compatible and backfill-free image search system are described. According to some embodiments, a backwards compatible training system trains a new embedding model to be backward compatible with the face embeddings (e.g., floating-point vectors) generated by a previous embedding model. In one embodiment, backwards compatible training uses a classifier of the previous embedding model as a form of constraint in the training of the new embedding model.

Abstract: A method for identifying a target, a non-transitory computer-readable storage medium, and an electronic device include: acquiring a first image and a second image, the first image and the second image each including a target to be determined; generating a prediction path based on the first image and the second image, both ends of the prediction path respectively corresponding to the first image and the second image; and performing validity determination on the prediction path and determining, according to a determination result, whether the targets to be determined in the first image and the second image are the same target to be determined.

Abstract: The present disclosure relates to a network optimization method and apparatus, an image processing method and apparatus, and a storage medium. The network optimization method includes: obtaining an image sample group; obtaining a first feature and a second feature of an image in the image sample group, and obtaining a first classification result by using the first feature of the image; performing feature exchange processing on an image pair in the image sample group to obtain a new image pair; obtaining a first loss value of the first classification result, a second loss value of the new image pair, and a third loss value of first features and second features of the new image pair in a preset manner; and adjusting parameters of a neural network at least according to the first loss value, the second loss value, and the third loss value until a preset requirement is met.

Abstract: The present disclosure discloses a rotary joint electromagnetic locking device and a rotary joint, the rotary joint electromagnetic locking device comprising a first electromagnetic sucker, a sliding plate, a rotary plate, a connecting shaft connecting the first electromagnetic sucker and the sliding plate, and a second electromagnetic sucker, different from the prior art. When the first electromagnetic sucker and the second electromagnetic sucker have no adsorption force, the sliding plate and the rotary plate can be locked to each other; when the second electromagnetic sucker has an adsorption force, the rotary plate or the first electromagnetic sucker is adsorbed by the second electromagnetic sucker, making the sliding plate and the rotary plate to be further locked; when the first electromagnetic sucker has an adsorption force, the sliding plate can be adsorbed to the first contact surface of the first electromagnetic sucker, making the sliding plate released from the rotary plate.

Abstract: A method for identifying a target, a non-transitory computer-readable storage medium, and an electronic device include: acquiring a first image and a second image, the first image and the second image each including a target to be determined; generating a prediction path based on the first image and the second image, both ends of the prediction path respectively corresponding to the first image and the second image; and performing validity determination on the prediction path and determining, according to a determination result, whether the targets to be determined in the first image and the second image are the same target to be determined.

Abstract: An end-effector is provided for use on a micro/nano manipulation device. The end-effector is comprised of: a micropump fluidly coupled to a microtube; a piezoelectric sensing structure disposed in the microtube; and a processing circuit electrically coupled to the sensing structure for determining the force of the fluid flowing through the microtube. The end-effector is a closed loop control-enabled micro/nano manipulation system.

Abstract: An integrated sensory actuator (10) which uses an electroactive polymer is provided. The sensory actuator is comprised of an actuating member (12) made of an ionic polymer-metal composite; a sensing member (14) made of a piezoelectric material; and an insulating member (16) interposed between the actuating member and the sensing member. The sensory actuator may further include a compensation circuit adapted to receive a sensed signal from the sensing member and an actuation signal from the actuating member and compensate the sensed signal for feedthrough coupling between the actuating member and the sensing member.

Abstract: An end-effector is provided for use on a micro/nano manipulation device. The end-effector is comprised of: a micropump fluidly coupled to a microtube; a piezoelectric sensing structure disposed in the microtube; and a processing circuit electrically coupled to the sensing structure for determining the force of the fluid flowing through the microtube. The end-effector is a closed loop control-enabled micro/nano manipulation system.

Abstract: An active micro-force sensor is provided for use on a micromanipulation device. The active micro-force sensor includes a cantilever structure having an actuator layer of piezoelectric material and a sensing layer of piezoelectric material bonded together. When an external force is exerted on the sensor, the sensor deforms and an applied force signal is recorded by the sensing layer. The applied force signal is then fed back to the actuating layer of the sensor via a servoed transfer function or servo controller, so that a counteracting deformation can be generated by the bending moment from the servoed actuating layer to quickly balance the deformation caused by the external micro-force.

Abstract: An improved microforce sensing system is provided. The microforce sensing system includes: a force sensor configured to detect a contact force exerted on a contact tip of the cantilever and operable to generate a signal indicative of the contact force, where the cantilever is constructed from a polyvinylidene fluoride material in the form of a substantially rectangular plate or other suitable shapes; and a processing circuit adapted to receive the signal from the force sensor and operable to determine the contact force exerted on the tip by integrating the signal from the force sensor.

Abstract: An active micro-force sensor is provided for use on a micromanipulation device. The active micro-force sensor includes a cantilever structure having an actuator layer of piezoelectric material and a sensing layer of piezoelectric material bonded together. When an external force is exerted on the sensor, the sensor deforms and an applied force signal is recorded by the sensing layer. The applied force signal is then fed back to the actuating layer of the sensor via a servoed transfer function or servo controller, so that a counteracting deformation can be generated by the bending moment from the servoed actuating layer to quickly balance the deformation caused by the external micro-force.

Abstract: An improved microforce sensing system is provided. The microforce sensing system includes: a force sensor configured to detect a contact force exerted on a contact tip of the cantilever and operable to generate a signal indicative of the contact force, where the cantilever is constructed from a polyvinylidene fluoride material in the form of a substantially rectangular plate or other suitable shapes; and a processing circuit adapted to receive the signal from the force sensor and operable to determine the contact force exerted on the tip by integrating the signal from the force sensor.