Abstract: An energy-saving electromagnetic brake includes a base, a first winding coil, a second winding coil, and a control circuit component. The first winding coil is disposed inside the base, wherein the first winding coil has a first resistance value. The second winding coil is disposed inside the base and is disposed around the first winding coil, wherein the second winding coil has a second resistance value, and the second resistance value is greater than the first resistance value. The control circuit component is disposed inside the base and is electrically connected to the first winding coil and the second winding coil. In a first period, the control circuit component drives the first winding coil. In a second period, the control circuit component simultaneously drives the first winding coil and the second winding coil, and the first winding coil and the second winding coil are connected in series.

Abstract: A control method of a consecutive convolution structure is provided. A first line, a second line of an input data and a first and second padding line can be adopted for applying Winograd convolution and generating a first line of an output data in an odd output layer. In addition, a first, a second and third line of an input data and a third padding line can be adopted for applying the Winograd convolution and generating a first line and a second line of an output data in an even output layer. The disclosed control method is proposed as integrating the Winograd convolution with a line interleaved process, so as to effectively reduce the array area of multiply accumulate circuits and power consumption thereof. In addition, a minimum latency of the convolutional neural network model can be achieved by employing the disclosed control method of the present invention.

Abstract: A method and a device for convolution computation are proposed. The method includes to receive input data, perform first convolution computation and second convolution computation by respectively using a first kernel and a second kernel according to the input data so as to generate plural computation results, and generate output data according to the computation results, where the size of the first kernel is different from that of the second kernel.

Abstract: A method and an apparatus for computation on a convolutional layer of a neural network are proposed. The apparatus includes an adder configured to receive a first sum of products, receive a pre-computed convolution bias of the convolutional layer, and perform accumulation on the first sum of products and the pre-computed convolution bias to generate an adder result of the convolutional layer, where the first sum of products is a sum of products of quantized input activation of the convolutional layer and quantized convolution weights of the convolutional layer, and where the pre-computed convolution bias is associated with a zero point of input activation of the convolutional layer and a zero point of output activation of the convolutional layer.

Abstract: An energy-saving electromagnetic brake includes a base, a first winding coil, a second winding coil, and a control circuit component. The first winding coil is disposed inside the base, wherein the first winding coil has a first resistance value. The second winding coil is disposed inside the base and is disposed around the first winding coil, wherein the second winding coil has a second resistance value, and the second resistance value is greater than the first resistance value. The control circuit component is disposed inside the base and is electrically connected to the first winding coil and the second winding coil. In a first period, the control circuit component drives the first winding coil. In a second period, the control circuit component simultaneously drives the first winding coil and the second winding coil, and the first winding coil and the second winding coil are connected in series.

Abstract: A vibration motor having an insulating housing, a bidirectional coil and a magnetic ring is provided. The insulating housing is of cylindrical shape and at least three position-limiting ribs are protruded from an internal surface thereof. Each position-limiting rib is extended parallel to a central axial line of the insulating housing. The insulating housing has a pair of end plates on two ends thereof. The end plates are made of insulating material. The bidirectional coil is accommodated in the insulating housing and spaced apart from the internal surface of the insulating housing. The magnetic ring is movably accommodated in the insulating housing, and the magnetic ring is arranged coaxially with the insulating housing to surround the bidirectional coil. The position-limiting ribs are disposed adjacent to the magnetic ring and surround an outer edge of the magnetic ring.

Abstract: The disclosure provides an image processing device. An image processing device includes a neural network circuit, a gain control circuit, and an image merge circuit. The neural network circuit is configured to receive an input image, and perform an image processing operation on the input image according to fixed parameters to output a first intermediate image. The gain control circuit, coupled to the neural network circuit, is configured to receive the first intermediate image from the neural network circuit, and multiply the first intermediate image by at least one amplitude gain to output a second intermediate image. The image merge circuit, coupled to the gain control circuit, configured to receive the second intermediate image and the input image, and combine the second intermediate image and the input image to obtain an output image. In addition, a method for image enhancement is also provided.

Abstract: A vibration motor having an insulating housing, a bidirectional coil and a magnetic ring is provided. The insulating housing is of cylindrical shape and at least three position-limiting ribs are protruded from an internal surface thereof. Each position-limiting rib is extended parallel to a central axial line of the insulating housing. The insulating housing has a pair of end plates on two ends thereof. The end plates are made of insulating material. The bidirectional coil is accommodated in the insulating housing and spaced apart from the internal surface of the insulating housing. The magnetic ring is movably accommodated in the insulating housing, and the magnetic ring is arranged coaxially with the insulating housing to surround the bidirectional coil. The position-limiting ribs are disposed adjacent to the magnetic ring and surround an outer edge of the magnetic ring.

Abstract: The disclosure provides an image processing device. An image processing device includes a neural network circuit, a gain control circuit, and an image merge circuit. The neural network circuit is configured to receive an input image, and perform an image processing operation on the input image according to fixed parameters to output a first intermediate image. The gain control circuit, coupled to the neural network circuit, is configured to receive the first intermediate image from the neural network circuit, and multiply the first intermediate image by at least one amplitude gain to output a second intermediate image. The image merge circuit, coupled to the gain control circuit, configured to receive the second intermediate image and the input image, and combine the second intermediate image and the input image to obtain an output image. In addition, a method for image enhancement is also provided.

Abstract: A display driver circuit, configured to drive a display panel, includes a time recording circuit, a storage circuit, and an output control circuit. The time recording circuit calculates a first time interval between a first vertical synchronous pulse and a second vertical synchronous pulse subsequent to the first vertical synchronous pulse. The storage circuit stores a display data corresponding to the first vertical synchronous pulse when the first vertical synchronous pulse is received by the display driver circuit. The output control circuit is coupled with the time recording circuit and the storage circuit. When the display driver circuit receives the second vertical synchronous pulse, the output control circuit outputs data pieces, generated by dividing the display data, to the display panel. The data pieces are outputted at second time intervals, and each of the second time intervals is positively correlated with the first time interval.

Abstract: A display driver circuit, configured to drive a display panel, includes a time recording circuit, a storage circuit, and an output control circuit. The time recording circuit calculates a first time interval between a first vertical synchronous pulse and a second vertical synchronous pulse subsequent to the first vertical synchronous pulse. The storage circuit stores a display data corresponding to the first vertical synchronous pulse when the first vertical synchronous pulse is received by the display driver circuit. The output control circuit is coupled with the time recording circuit and the storage circuit. When the display driver circuit receives the second vertical synchronous pulse, the output control circuit outputs data pieces, generated by dividing the display data, to the display panel. The data pieces are outputted at second time intervals, and each of the second time intervals is positively correlated with the first time interval.

Abstract: A video system includes a scene analyzing circuit, configured to perform a scene analysis operation on a video signal and output a scene analysis result; and a video processing module, comprising a first video processing circuit, configured to perform a first video operation on the video signal and output a first processing result; and a second video processing circuit, configured to perform a second video operation on the first processing result according to the scene analysis result.

Abstract: A video system includes a scene analyzing circuit, configured to perform a scene analysis operation on a video signal and output a scene analysis result; and a video processing module, comprising a first video processing circuit, configured to perform a first video operation on the video signal and output a first processing result; and a second video processing circuit, configured to perform a second video operation on the first processing result according to the scene analysis result.

Abstract: The disclosure proposes a method and a system for 360-degree video playback. The method is applicable to a video playback system and includes at least the following steps. A current frame of a 360-degree video having a sequence of frames is received. Candidate objects in the current frame are detected, and a main object is selected from the candidate objects by using a selector recurrent neural network (RNN) model based on information of the candidate objects in the current frame. A viewing angle corresponding to the current frame is computed by using a regressor RNN model based on the main object in the current frame. The current frame is displayed on the screen according to the viewing angle.

Abstract: An apparatus for motion compensated noise reduction for input images is provided. The motion estimation and motion compensation circuit performs a motion estimation operation and a motion compensation operation on a current image and a previous image to obtain a first patch. The block matching operation circuit performs a block matching operation on the current image and the previous image to obtain a second patch. The motion detection circuit performs a motion detection operation on a target patch according to the first patch and the second patch to output a set of third patches. The current image includes the target patch. The noise reduction circuit performs a noise reduction operation on the set of third patches according to a threshold curve, so as to generate the target patch that the noise is reduced. A method for motion compensated noise reduction for input images is also provided.

Abstract: An apparatus including a processor and a non-transitory storage medium is provided. The non-transitory storage medium includes a program code executable by the processor. The apparatus is configured to execute the program code with the processor for processing a source image to generate a target image.

Abstract: An apparatus including a processor and a non-transitory storage medium is provided. The non-transitory storage medium includes a program code executable by the processor. The apparatus is configured to execute the program code with the processor for processing a source image to generate a target image.

Abstract: The disclosure proposes a method and a system for 360-degree video playback. The method is applicable to a video playback system and includes at least the following steps. A current frame of a 360-degree video having a sequence of frames is received. Candidate objects in the current frame are detected, and a main object is selected from the candidate objects by using a selector recurrent neural network (RNN) model based on information of the candidate objects in the current frame. A viewing angle corresponding to the current frame is computed by using a regressor RNN model based on the main object in the current frame. The current frame is displayed on the screen according to the viewing angle.

Abstract: An apparatus for processing a source image to generate a target image is provided. The target image includes a plurality of target pictures. The apparatus includes a first image processing circuit and a second image processing circuit. The first image processing circuit separates the source image into a first image including a first part of the source image and a second image including a second part of the source image. The second image processing circuit processes the first image to obtain a plurality of target first pictures according to a first setting, process the second image to obtain a plurality of target second pictures according to a second setting, and combine each of the target first pictures with a corresponding one of the target second pictures to obtain the plurality of target pictures. A method for processing a source image to generate a target image is also provided.

Abstract: An apparatus for motion compensated noise reduction for input images is provided. The motion estimation and motion compensation circuit performs a motion estimation operation and a motion compensation operation on a current image and a previous image to obtain a first patch. The block matching operation circuit performs a block matching operation on the current image and the previous image to obtain a second patch. The motion detection circuit performs a motion detection operation on a target patch according to the first patch and the second patch to output a set of third patches. The current image includes the target patch. The noise reduction circuit performs a noise reduction operation on the set of third patches according to a threshold curve, so as to generate the target patch that the noise is reduced. A method for motion compensated noise reduction for input images is also provided.