Abstract: A button mechanism is provided, including a frame, a motor disposed on the frame, a rotary member connected to the motor, a protrusion disposed on the rotary member, and a movable member hinged to the frame. When the movable member moves relative to the frame in the first direction, the protrusion contacts the movable member, and the rotary member is driven by the motor to rotate, whereby the movable member is pushed back by the protrusion so that it moves in the second direction, which is the opposite of the first direction.

Abstract: An automated molecular operating system includes at least one centrifuge tube carrying module, a transport module, a plurality of temperature control modules, a capping module, a magnetic field module and an automated processing module. The automated processing module is electrically connected to the transport module, the temperature control modules, the capping module and the magnetic field module, and controls the transport module to move the centrifuge tube carrying module, so that a centrifuge tube contained in the centrifuge tube carrying module makes a reaction in the temperature control modules, and the magnetic field module or the capping module is provided to the centrifuge tube according to requirements, such that a specimen in the centrifuge tube can be automatically subjected to nucleic acid extraction, nucleic acid amplification, primer labeling, reverse transcription or a combination thereof, thereby reducing manual operation errors and increasing the ease of operation.

Abstract: A semiconductor structure includes a substrate, an isolation structure disposed in the substrate, and a hybrid structure disposed over the isolation structure. The hybrid structure is substantially conformal with respect to a profile of the isolation structure. The hybrid structure includes an oxide component, a nitride component surrounding the oxide component, and a first polysilicon component alongside the nitride component. The nitride component includes a first upper surface closed to the first polysilicon component, and a second upper surface distal to the first polysilicon component. The second upper surface is lower than the first upper surface.

Abstract: A semiconductor device is disclosed. The device includes a source/drain feature formed over a substrate. A dielectric layer formed over the source/drain feature. A contact trench formed through the dielectric layer to expose the source/drain feature. A titanium nitride (TiN) layer deposited in the contact trench and a cobalt layer deposited over the TiN layer in the contact trench.

Abstract: An organic-waste-recycling apparatus for fast formation of compost includes a processing barrel, a grinding device, a control device, a stirring device, and a guiding device. The processing barrel has a first feeding port and a second feeding port, a mixing room, and an output. To-be-processed objects are fed into the mixing room through the first feeding port or the second feeding port. The grinding device is between the first feeding port and the mixing room. The control device adjusts the temperature and the humidity in the mixing room. The stirring device rolls and blends the to-be-processed objects in the mixing room so that the to-be-processed objects are heated evenly. The guiding device guides the water vapor generated in the mixing room back to the mixing room. The apparatus decomposes and turns organic waste into organic carbon fertilizer that has economic value in an effective and efficient way.

Abstract: Disclosed herein is a method to disentangle linear-encoded facial semantics from facial images without external supervision. The method uses linear regression and sparse representation learning concepts to make the disentangled latent representations easily interpreted and manipulated. Generated facial images are decomposed into multiple semantic features and latent representations are extracted to capture interpretable facial semantics. The semantic features may be manipulated to synthesize photorealistic facial images by sampling along vectors representing the semantic features, thereby changing the associate semantics.

Abstract: A device, suitable for a V2X (Vehicle-to-everything) system, is disclosed. The device includes a wireless transceiver and a processor. The wireless transceiver is configured to receive a first collective perception message from a transmitting device of the V2X system. The first collective perception message comprises a first object age of a perceived object of the first collective perception message calculated by the transmitting device. The processor is coupled to the wireless transceiver, and the processor is configured to calculate a second object age of the perceived object of a second collective perception message. The second object age comprises a processing delay of the device and the first object age.

Abstract: Disclosed herein is a system and method for data augmentation for general object recognition which preserves the class identity of the augmented data. The system comprises an image recognition network an image generation network that take as input ground truth images and classes respectively and which generates a predicted class and an augmented image. A discriminator evaluates the predicted class and augmented image and provides feedback to the image recognition network and the image generation network.

Abstract: Disclosed here are various techniques for improving the testing and training of datasets comprising sequences of skeletal representations performing various actions. The dataset can be denoised by applying various techniques to determine noisy frames within each sequence and eliminating the sequences from the dataset when the number of noisy frames in the sequence is too large. In addition, the dataset may be augmented by various data augmentation techniques to manipulate the skeletal representations, after denoising.

Abstract: Disclosed here are various techniques for improving the testing and training of datasets comprising sequences of skeletal representations performing various actions. The dataset can be denoised by applying various techniques to determine noisy frames within each sequence and eliminating the sequences from the dataset when the number of noisy frames in the sequence is too large. In addition, the dataset may be augmented by various data augmentation techniques to manipulate the skeletal representations, after denoising.

Abstract: Disclosed herein is a method to disentangle linear-encoded facial semantics from facial images without external supervision. The method uses linear regression and sparse representation learning concepts to make the disentangled latent representations easily interpreted and manipulated. Generated facial images are decomposed into multiple semantic features and latent representations are extracted to capture interpretable facial semantics. The semantic features may be manipulated to synthesize photorealistic facial images by sampling along vectors representing the semantic features, thereby changing the associate semantics.

Abstract: A device, suitable for a V2X (Vehicle-to-everything) system, is disclosed. The device includes a wireless transceiver and a processor. The wireless transceiver is configured to receive a first collective perception message from a transmitting device of the V2X system. The first collective perception message comprises a first object age of a perceived object of the first collective perception message calculated by the transmitting device. The processor is coupled to the wireless transceiver, and the processor is configured to calculate a second object age of the perceived object of a second collective perception message. The second object age comprises a processing delay of the device and the first object age.

Abstract: An electronic device used in a method for expanding image depth obtains first images by a first sensor, the first images comprising depth information. The electronic device obtains second images by a second sensor, the second images comprising gradient information, and the first images correspond to the second images. The electronic device determines the pixels in the first images which contain expandable content according to the gradient information of the second images, applies expansion accordingly to the pixels in the first images to generate third images, and generate target depth maps according to the gradient information of the second images and the depth information of the third images.

Abstract: The present invention relates to a traffic situation detecting method. The method includes steps of acquiring a current location, a speed of movement and a direction of movement for a ground vehicle; using the current location as a starting point and drawing a detecting scope on a digital map toward the direction of movement; varying a size of the detecting scope in adaptive to an interval of the speed to which the speed of movement belongs; accessing a traffic server to retrieve a traffic event and determining whether the traffic event is situated within the detecting scope; and marking the traffic event located within the detecting scope on a digital map.

Abstract: The present disclosure relates to a use of an ADAM9 inhibitor in manufacture of an immunomodulator. The immunomodulator can change the distribution of immune cells in a tumor and increase the infiltration of the immune cells, thereby increasing the efficacy of cancer immunotherapy. Furthermore, the immunomodulator can be used to enhance an effectiveness of a cancer immunotherapy, and the ADAM9 inhibitor can be a compound of Formula (I), or a pharmaceutically acceptable salt or stereoisomer thereof.

Abstract: An electronic device used in a method for expanding image depth obtains first images by a first sensor, the first images comprising depth information. The electronic device obtains second images by a second sensor, the second images comprising gradient information, and the first images correspond to the second images. The electronic device determines the pixels in the first images which contain expandable content according to the gradient information of the second images, applies expansion accordingly to the pixels in the first images to generate third images, and generate target depth maps according to the gradient information of the second images and the depth information of the third images.

Abstract: The present invention relates to a traffic situation detecting method. The method includes steps of acquiring a current location, a speed of movement and a direction of movement for a ground vehicle; using the current location as a starting point and drawing a detecting scope on a digital map toward the direction of movement; varying a size of the detecting scope in adaptive to an interval of the speed to which the speed of movement belongs; accessing a traffic server to retrieve a traffic event and determining whether the traffic event is situated within the detecting scope; and marking the traffic event located within the detecting scope on a digital map.