The present invention relates to the technical field of identification on adulterated meat, and in particular, to a method for identifying raw meat and high-quality fake meat based on a gradual linear array change of a component. The present invention spatially characterizes changing rules of featured components in the meat with the utilization of sensitivities of the visible/near-infrared spectral signals to changes of the components in the meat and the advantage that spectral scanning can acquire optical signals of the samples spatially and consecutively, further constructs the identification model according to differences in components and spectra of a region of interest in the hyperspectral image by taking a derivative for characterizing rates of change of the featured components.

A thickness estimating apparatus includes a transfer robot, a light source, a camera, a memory and a controller. The memory stores a thickness predicting model generated based on a data set including a thickness of at least one of a test wafer corresponding to the wafer or a test element layer formed on the test wafer, and the thickness predicting model being trained to minimize a loss function of the data set. The controller applies pixel data, which is acquired from at least one pixel selected from a plurality of pixels included in a captured image, to the thickness predicting model, to predict a thickness of at least one of the wafer or an element layer formed on the wafer in a position corresponding to a position of the selected pixel.

A learning data generation device that can generate learning data suitable for learning of an identification model. The learning data generation device has a function of cutting out part of first image data as second image data, a function of generating a two-dimensional graphic corresponding to the area of the second image data and representing a pseudo defect, a function of generating third image data by combining the second image data and the two-dimensional graphic, and a function of assigning a label corresponding to the two-dimensional graphic to the third image data. By using the third image data for learning of the identification model, a highly accurate identification model can be generated.

A method for automatically reconstituting the architecture, along a reinforcing axis, of the reinforcement of a composite material, includes acquiring images of the reinforcement of the composite material, each image being acquired along a section plane perpendicular to the reinforcing axis; for each image acquired, detecting, using a neural network, barycentre and/or the circumference of each section of the reinforcing thread; for at least one acquired reference image, assigning a tag corresponding to a reinforcing thread, to each detected barycentre or circumference; for each other acquired image, assigning, to each detected barycentre and/or each detected circumference, the tag of the corresponding barycentre in the acquired reference image; reconstituting the architecture of each reinforcing thread from each detected barycentre and/or circumference having the tag of the reinforcing thread and the position on the reinforcing axis associated with the acquired image on which the barycentre and/or the circumference has been detected.

There is provided a method of examination of a semiconductor specimen and a system thereof. The method comprises: using a trained Deep Neural Network (DNN) to process a fabrication process (FP) sample, wherein the FP sample comprises first FP image(s) received from first examination modality(s) and second FP image(s) received from second examination modality(s) which differs from the first examination modality(s), and wherein the trained DNN processes the first FP image(s) separately from the second FP image(s); and further processing by the trained DNN the results of such separate processing to obtain examination-related data specific for the given application and characterizing at least one of the processed FP images. When the FP sample further comprises numeric data associated with the FP image(s), the method further comprises processing by the trained DNN at least part of the numeric data separately from processing the first and the second FP images.

An image-based acceptance learning device (2) learns a result of determination as to whether a planar object (1) is acceptable or defective based on at least one of a three-dimensional shape or a color on a surface of the planar object (1). The device (2) includes a surface image receiver (3) to receive an inputted two-dimensional data being image data of the surface of the planar object (1), a determination information receiver (4) to receive an inputted determination information indicating the result of determination as to whether the planar object (1) corresponding to the two-dimensional data is acceptable or defective, and a learner (5) to learn, based on the two-dimensional data and the determination information, a relevant area (1R) including the three-dimensional shape or the color on the surface in the two-dimensional data. The relevant area is a basis for the determination information.

Methods and systems for monitoring the quality of a semiconductor measurement in a targeted manner are presented herein. Rather than relying on one or more general indices to determine overall measurement quality, one or more targeted measurement quality indicators are determined. Each targeted measurement quality indicator provides insight into whether a specific operational issue is adversely affecting measurement quality. In this manner, the one or more targeted measurement quality indicators not only highlight deficient measurements, but also provide insight into specific operational issues contributing to measurement deficiency. In some embodiments, values of one or more targeted measurement quality indicators are determined based on features extracted from measurement data. In some embodiments, values of one or more targeted measurement quality indicators are determined based on features extracted from one or more indications of a comparison between measurement data and corresponding measurement data simulated by a trained measurement model.

A terahertz scanner for detecting irregularities, such as chemical or structural variations, in a sample and methods of use thereof are described. The described terahertz scanner and algorithms allow for direct, high-sensitivity, high-throughput, and non-invasive detection of irregularities that range from chemical contaminant to material defects in a variety of substrates and settings.

A system and methods for OCD metrology are provided including receiving training data for training an OCD machine learning (ML) model, including multiple pairs of corresponding sets of scatterometric data and reference parameters. For each of the pairs, one or more corresponding outlier metrics are by calculated and corresponding outlier thresholds are applied whether a given pair is an outlier pair. The OCD MIL model is then trained with the training data less the outlier pairs.

A system and methods for OCD metrology are provided including receiving multiple first sets of scatterometric data, dividing each set into k sub-vectors, and training, in a self-supervised manner, k2 auto-encoder neural networks that map each of the k sub-vectors to each other. Subsequently multiple respective sets of reference parameters and multiple corresponding second sets of scatterometric data are received and a transfer neural network (NN) is trained. Initial layers include a parallel arrangement of the k2 encoder neural networks. Target output of the transfer NN training is set to the multiple sets of reference parameters and feature input is set to the multiple corresponding second sets of scatterometric data, such that the transfer NN is trained to estimate new wafer pattern parameters from subsequently measured sets of scatterometric data.