A method of analysing the quality of a welding area of an absorbent sanitary article is disclosed. During a learning step, a plurality of welding operations are performed both with a sufficient quality and with an insufficient quality, and the welding area generated for each welding operation is monitored via a camera. During a training step, the pixel data of the welding areas monitored during the learning step is processed for training a classifier configured to estimate a welding quality as a function of respective pixel data of a respective welding area. Accordingly, during a normal welding operating step, the welding quality may be estimate via the classifier, thereby improving the environmental sustainability of the production process.

A machine learning system may automatically produce classifier algorithms and configuration parameters by selecting them into a set of predetermined unitary algorithms and associated parametrization values. Multiple digital representations of input object items may be produced by varying the position and orientation of the object to be classified and/or of the sensor to capture a digital representation of the object, and/or by varying a physical environment parameter which changes the digital representation capture of the object by the sensor. A robot arm or a conveyor may vary the object and/or the sensor positions and orientations. The machine learning system may employ genetic programming to facilitate the production of classifiers suitable for the classification of multiple digital representations of input object items. The machine learning system may automatically generate reference template signals as configuration parameters for the unitary algorithms to facilitate the production of classifiers suitable for the classification of multiple digital representations of input object items.

An information processing device is an information processing device including a processor. The processor obtains a detection result of a first detector for detecting a first target in first sensing data; and based on the detection result of the first detector, determines a setting of processing by a second detector for detecting a second target in second sensing data next in an order after the first sensing data, the second target being different from the first target.

A system and method for quantifying Alzheimer's disease (AD) risk using one or more interferometric micro-Doppler radars (IMDRs) and deep learning artificial intelligence to distinguish between cognitively unimpaired individuals and persons with AD based on gait analysis. The system utilizes IMDR to capture signals from both radial and transversal movement in three-dimensional space to further increase the accuracy for human gait estimation. New deep learning technologies are designed to complement traditional machine learning involving separate feature extraction followed-up with classification to process radar signature from different views including side, front, depth, limbs, and whole body where some motion patterns are not easily describable. The disclosed cross-talk deep model is the first to apply deep learning to learn IMDR signatures from two perpendicular directions jointly from both healthy and unhealthy individuals. Decision fusion is used to integrate classification results from feature-based classifier and deep learning AI to reach optimal decision.

In accordance with various embodiments, a method (100) for the computer-aided recognition of a transport container (402) being empty can comprise: determining (101) one or more than one segment of image data of the transport container (402) using depth information assigned to the image data; assigning (103) the image data to one from a plurality of classes, of which a first class represents the transport container (402) being empty, and a second class represents the transport container (402) being non-empty, wherein the one or more than one segment is not taken into account in the assigning; outputting (105) a signal which represents the filtered image data being assigned to the second class.

In one embodiment, a method includes accessing a batch B of a plurality of images, wherein each image in the batch is part of a training set of images used to train a vision transformer comprising a plurality of attention heads. The method further includes determining, for each attention head A, a similarity between (1) the output of the attention head evaluated using each image in the batch and the (2) output of each attention head evaluated using each image in the batch. The method further includes determining, based on the determined similarities, an importance score for each attention head; and pruning, based on the importance scores, one or more attention heads from the vision transformer.

Methods and systems for adjusting formulation of pigments for dye sublimation are disclosed. In one embodiment, the method includes receiving, by a processor, a selection of the substrate and the image to be dye sublimated into the substrate. The method also includes determining, by the processor, an amount of each pigment of sublimation ink to be used to print the image on a sheet based on a characteristic of the substrate. The method further includes printing, by a dye sublimation apparatus, the image on the sheet using the amount of each pigment of sublimation ink and infusing, by the dye sublimation apparatus, the image from the sheet into the substrate.

An image analysis device has: an image analysis circuitry to analyze images input from each of cameras using instances of an image analysis program including a learned neural network model for object detection and learned neural network models for object recognition; inference processors to perform inference processes in the learned neural network model for object detection and each learned neural network model for object recognition; and a processor assignment circuitry to assign, from the inference processors, inference processors to be used for the inference process in the learned neural network model for object detection and the inference process in each learned neural network model for object recognition, based on inference time and frequency of use required for the inference process in each of the learned neural network models for object detection and object recognition included in each instance of the image analysis program.

Provided are a face feature point detection method, applied to an image processing device, where the image processing device stores a feature area detection model and a feature point detection model. The method includes: preprocessing a face image to be detected to obtain a preprocessed target face image; performing feature point extraction on the target face image according to the feature area detection model and the feature point detection model to obtain a target feature point coordinate located within a face feature area in the target face image; and performing coordinate transformation on the target feature point coordinate to obtain a face feature point coordinate corresponding to the face image to be detected. Further provided are a face feature point detection device, an equipment and a storage medium.

Generally, the present disclosure is directed to novel machine-learned classification models that operate with hard attention to make discrete attention actions. The present disclosure also provides a self-supervised pre-training procedure that initializes the model to a state with more frequent rewards. Given only the ground truth classification labels for a set of training inputs (e.g., images), the proposed models are able to learn a policy over discrete attention locations that identifies certain portions of the input (e.g., patches of the images) that are relevant to the classification. In such fashion, the models are able to provide high accuracy classifications while also providing an explicit and interpretable basis for the decision.