Abstract: An optical system affixed to an electronic apparatus is provided, including a first optical module, a second optical module, and a third optical module. The first optical module is configured to adjust the moving direction of a first light from a first moving direction to a second moving direction, wherein the first moving direction is not parallel to the second moving direction. The second optical module is configured to receive the first light moving in the second moving direction. The first light reaches the third optical module via the first optical module and the second optical module in sequence. The third optical module includes a first photoelectric converter configured to transform the first light into a first image signal.

Abstract: A classification model building apparatus and a classification model building method thereof are provided. The classification model building apparatus introduces a clustering algorithm to assist in training a deep learning model for classification and takes a sum of a clustering loss function, a center concentration loss function and a classification loss function multiplied by different weights, respectively, as a total loss function for training the deep learning model. Based on the total loss function, the classification model building apparatus adjusts parameters of the deep learning model through a backpropagation algorithm to build a classification model.

Abstract: An action evaluation model building apparatus and an action evaluation model building method thereof are provided. The action evaluation model building apparatus stores a plurality of raw data sets and a plurality of standard action labels corresponding thereto. Based on machine learning algorithms, the action evaluation model building apparatus computes the raw data sets and performs a supervised learning to build a feature vector creation model and a classifier model. The action evaluation model building apparatus determines a representation action feature vector of each standard action label by randomly generating a plurality of action feature vectors and inputting them into the classifier model. The action evaluation model building apparatus builds an action evaluation model based on the feature vector creation model, the classifier model and the representation action feature vectors.

Abstract: A motion recognition apparatus extracts a plurality of first motion feature vectors within a period of time, and extracts a plurality of second motion feature vectors within another period of time closely next to the period of time. The apparatus calculates and sequences the distances between each of the plurality of first motion feature vectors and the plurality of cluster centroids to generate a corresponding first cluster centroid sequence respectively, and calculates and sequences the distances between each of the plurality of second motion feature vectors and the plurality of cluster centroids to generate a corresponding second cluster centroid sequence respectively. The apparatus recognizes a first segment motion, a second segment motion, and a whole motion based on the number of appearances of the cluster centroid in the first cluster centroid sequences, the second cluster centroid sequences and both of them respectively.

Abstract: A motion recognition apparatus extracts a plurality of first motion feature vectors within a period of time, and extracts a plurality of second motion feature vectors within another period of time closely next to the period of time. The apparatus calculates and sequences the distances between each of the plurality of first motion feature vectors and the plurality of cluster centroids to generate a corresponding first cluster centroid sequence respectively, and calculates and sequences the distances between each of the plurality of second motion feature vectors and the plurality of cluster centroids to generate a corresponding second cluster centroid sequence respectively. The apparatus recognizes a first segment motion, a second segment motion, and a whole motion based on the number of appearances of the cluster centroid in the first cluster centroid sequences, the second cluster centroid sequences and both of them respectively.

Abstract: A classification model building apparatus and a classification model building method thereof are provided. The classification model building apparatus introduces a clustering algorithm to assist in training a deep learning model for classification and takes a sum of a clustering loss function, a center concentration loss function and a classification loss function multiplied by different weights, respectively, as a total loss function for training the deep learning model. Based on the total loss function, the classification model building apparatus adjusts parameters of the deep learning model through a backpropagation algorithm to build a classification model.

Abstract: A model training system, an operating method of the model training system and a portable electronic device are provided. The operating method is implemented by a processor. The model training system includes a first model and at least one second model. The processor receives an inputting data. The processor obtains a model difference between the first model and the second model according to the inputting data. The processor obtains an outputting accuracy of the second model according to the inputting data. The processor updates at least one parameter of the second model according to the outputting accuracy and the model difference.

Abstract: An action evaluation model building apparatus and an action evaluation model building method thereof are provided. The action evaluation model building apparatus stores a plurality of raw data sets and a plurality of standard action labels corresponding thereto. Based on machine learning algorithms, the action evaluation model building apparatus computes the raw data sets and performs a supervised learning to build a feature vector creation model and a classifier model. The action evaluation model building apparatus determines a representation action feature vector of each standard action label by randomly generating a plurality of action feature vectors and inputting them into the classifier model. The action evaluation model building apparatus builds an action evaluation model based on the feature vector creation model, the classifier model and the representation action feature vectors.

Abstract: A semiconductor structure includes a stacked metal oxide layer on a substrate, wherein the stacked metal oxide layer includes a first metal oxide layer, a second metal oxide layer, and a third metal oxide layer from top to bottom, and the energy bandgap of the second metal oxide layer is lower than the energy bandgap of the first metal oxide layer and that of the third metal oxide layer. The semiconductor structure includes a metal oxide layer on a substrate, wherein the energy bandgap of the metal oxide layer changes along a direction perpendicular to the surface of the substrate. The present invention also provides a semiconductor process forming said semiconductor structure.

Abstract: An identification method includes: sensing movement data; capturing multiple feature data from the movement data; cutting the first feature data into a plurality of first feature segments, dividing the first feature segments into a plurality of first feature groups, and calculating multiple first similarity parameters of the first feature groups respectively corresponding to a plurality of channels; making the first feature groups correspond to the channels according to the first similarity parameters; simplifying the first feature groups corresponding to the channels respectively by a convolution algorithm to obtain a plurality of first convolution results corresponding to the first feature groups; simplifying the first convolution results corresponding to the first feature groups respectively by a pooling algorithm to obtain multiple first pooling results corresponding to the first feature groups; and combining the first pooling results corresponding to the first feature groups to generate a first feature m

Abstract: An identification method includes: sensing movement data; capturing multiple feature data from the movement data; cutting the first feature data into a plurality of first feature segments, dividing the first feature segments into a plurality of first feature groups, and calculating multiple first similarity parameters of the first feature groups respectively corresponding to a plurality of channels; making the first feature groups correspond to the channels according to the first similarity parameters; simplifying the first feature groups corresponding to the channels respectively by a convolution algorithm to obtain a plurality of first convolution results corresponding to the first feature groups; simplifying the first convolution results corresponding to the first feature groups respectively by a pooling algorithm to obtain multiple first pooling results corresponding to the first feature groups; and combining the first pooling results corresponding to the first feature groups to generate a first feature m

Abstract: A semiconductor structure includes a stacked metal oxide layer on a substrate, wherein the stacked metal oxide layer includes a first metal oxide layer, a second metal oxide layer, and a third metal oxide layer from top to bottom, and the energy bandgap of the second metal oxide layer is lower than the energy bandgap of the first metal oxide layer and that of the third metal oxide layer. The semiconductor structure includes a metal oxide layer on a substrate, wherein the energy bandgap of the metal oxide layer changes along a direction perpendicular to the surface of the substrate. The present invention also provides a semiconductor process forming said semiconductor structure.

Abstract: A semiconductor structure includes a stacked metal oxide layer on a substrate, wherein the stacked metal oxide layer includes a first metal oxide layer, a second metal oxide layer, and a third metal oxide layer from top to bottom, and the energy bandgap of the second metal oxide layer is lower than the energy bandgap of the first metal oxide layer and that of the third metal oxide layer. The semiconductor structure includes a metal oxide layer on a substrate, wherein the energy bandgap of the metal oxide layer changes along a direction perpendicular to the surface of the substrate. The present invention also provides a semiconductor process forming said semiconductor structure.

Abstract: An ALD method includes providing a substrate in an ALD reactor, performing a pre-ALD treatment to the substrate in the ALD reactor, and performing one or more ALD cycles to form a dielectric layer on the substrate in the ALD reactor. The pre-ALD treatment includes providing a hydroxylating agent to the substrate in a first duration, and providing a precursor to the substrate in a second duration. Each of the ALD cycles includes providing the hydroxylating agent to the substrate in a third duration, and providing the precursor to the substrate in a fourth duration. The first duration is longer than the third duration.

Abstract: A semiconductor structure includes a stacked metal oxide layer on a substrate, wherein the stacked metal oxide layer includes a first metal oxide layer, a second metal oxide layer, and a third metal oxide layer from top to bottom, and the energy bandgap of the second metal oxide layer is lower than the energy bandgap of the first metal oxide layer and that of the third metal oxide layer. The semiconductor structure includes a metal oxide layer on a substrate, wherein the energy bandgap of the metal oxide layer changes along a direction perpendicular to the surface of the substrate. The present invention also provides a semiconductor process forming said semiconductor structure.

Abstract: A motion estimation method for video compression comprises the following steps. First, an initial simplex comprising three points is determined based on motion vectors in blocks of a current frame and a previous frame, and a point having a largest function value among the three points is replaced with a point having a smaller function value to form a simplex. The replacement is repeated until two points of the three points of the simplex converge to a same point. The iteration is performed by downhill simplex search including operations of reflection, expansion, contraction and shrinkage to find a point for replacement. The motion estimation method for video compression can also use multi-reference frames.

Abstract: A semiconductor process having a dielectric layer including metal oxide is provided. The semiconductor process includes: A substrate is provided. A dielectric layer including metal oxide is formed on the substrate, wherein the dielectric layer has a plurality of oxygen-related vacancies. A first oxygen-importing process is performed to fill the oxygen-related vacancies with oxygen. Otherwise, three MOS transistor processes are also provided, each of which has a gate dielectric layer including a high dielectric constant, and a first oxygen-importing process is performed to fill the oxygen-related vacancies with oxygen.

Abstract: A method for fabricating high-k dielectric layer is disclosed. The method includes the steps of: providing a substrate; and forming a plurality of high-k dielectric layers by using a plurality of reacting gases to perform a plurality of process stages on the surface of the substrate, wherein at least one of the reacting gases comprises different flow rate in the fabrication stages.

Abstract: A semiconductor process having a dielectric layer including metal oxide is provided. The semiconductor process includes: A substrate is provided. A dielectric layer including metal oxide is formed on the substrate, wherein the dielectric layer has a plurality of oxygen-related vacancies. A first oxygen-importing process is performed to fill the oxygen-related vacancies with oxygen. Otherwise, three MOS transistor processes are also provided, each of which has a gate dielectric layer including a high dielectric constant, and a first oxygen-importing process is performed to fill the oxygen-related vacancies with oxygen.

Abstract: A method for switching decoupled plasma nitridation (DPN) processes of different doses, which is able to decrease the switching time, is provided. According to the method, a dummy wafer is inserted into a chamber, a process gas introduced is ignited into plasma, and then a DPN doping process of the next dose is performed on the dummy wafer. The nitrogen concentration of the chamber is thus adjusted rapidly to switch to the DPN process of the next dose. In addition, after several cycles of the above steps are repeated, a dummy wafer is inserted into the chamber, and a complete DPN process of the next dose is performed on the dummy wafer. This process is performed several times before switching to the next DPN process.