Abstract: This invention relates to a data dimension reduction method based on maximizing a ratio sum for linear discriminant analysis, which belongs to the fields of image classification and pattern recognition. It includes constructing a data matrix, a label vector and a label matrix; calculating a within-class covariance matrix and a between-class covariance matrix; constructing the optimization problem based on maximizing the ratio sum for the linear discriminant analysis; using the alternating direction method of multipliers to obtain the projection matrix which can maximize an objective function. This invention establishes the objective function based on maximizing the ratio sum for the linear discriminant analysis to avoid the problem that the traditional linear discriminant analysis tends to select features with small variances and weak discriminating ability. It can select features which are more conducive to classification.

Abstract: The invention provides a global and local feature reconstruction network-based medical image segmentation method, device, apparatus and a computer storage medium. In the invention, a global feature reconstruction GFR module introduces semantic information from high-level into low-level features through a global descriptor, to eliminate a semantic gap between features in different levels, and a feature map is reconstructed by using a cross-level global descriptor, to implement long-distance feature dependency modeling. A local feature reconstruction LFR module implements dynamic upsampling of features based on the guidance of a low-stage feature map.

Abstract: In a method of segmenting a tubular feature in an image, a sequence of overlapping portions of the image are segmented using a trained model. The overlapping portions are positioned along the length of the tubular feature and combined to determine a segmentation of the tubular feature.

Abstract: A stenosis assessment method and device based on the intracranial digital subtraction angiographic (DSA) imaging, including acquiring the intracranial DSA imaging and extracting two planar images containing the target blood vessel from the DSA imaging, wherein the two planar images have different shooting angles. According to the two planar images, a 3D model of the target vessel is established. Based on the established 3D model of the target vessel and the DSA imaging, the hemodynamic simulation of the target vessel is performed. The disclosure realizes the functional assessment of intracranial vascular stenosis, improves the diagnostic accuracy, and provides certain assistance for neurologists to determine intervention means. The disclosure of noninvasive FFR technology in the assessment of intracranial vascular stenosis can only rely on angiography for functional assessment, saving the medical examination cost of patients. It has more convenient operation and higher repeatability.

Abstract: Methods for predicting a yield of fruit growing in an agricultural plot are provided. At a first time, a first plurality of images of a canopy of the agricultural plot is obtained from an aerial view of the canopy of the agricultural plot. From the first plurality of images, a first number of detectable fruit is estimated. At a second time, a second plurality of images of the canopy of the agricultural plot is obtained from the aerial view of the canopy of the agricultural plot. From the second plurality of images, a second number of detectable fruit is estimated. Using at least the first number of detectable fruit and the second number of detectable fruit and agricultural plot information, predict the yield of fruit from the agricultural plot.

Abstract: An image processing method, including: obtaining an original image, the original image including biological features of a to-be-detected object, fusing a global frequency domain map corresponding to the original image with the original image to obtain a global fusion image, performing living body detection on the global fusion image to obtain a first detection result, determining that the original image does not pass the living body detection when the first detection result indicates that the original image belongs to a screen reproduced image, obtaining, when the original image does not belong to the screen reproduced image, a biological feature image based on the biological features, fusing a local frequency domain map with the biological feature image to obtain a local fusion image, performing living body detection on the local fusion image to obtain a second detection result, and determining a living body detection result corresponding to the original image.

Abstract: Methods and systems described herein relate to authenticating users-based on generating images that have modified features. More specifically, the methods and systems generate these images by processing existing training data to identify a common feature in existing photos that may be modified, and then modifying that feature with the use of generative adversarial networks.

Abstract: Quality improvement processing to which a learning result is applied can be performed appropriately even if the signal characteristics of an input signal are different from the signal characteristics of student data at the time of learning. A quality improvement processing unit applies quality improvement processing to an input signal to obtain an output signal using a learning result obtained using a signal of a first domain having first signal characteristics as student data. A domain conversion unit converts the input signal to a signal of the first domain and sends the same to a quality improvement processing unit when the input signal is a signal of a second domain having second signal characteristics different from the first signal characteristics.

Abstract: Computerized systems are provided for applying data indicative of a personal style to a feature of a user represented in one or more images based on determining or estimating the personal style. In operation, embodiments can receive a first image of a first user that indicates the personal style of the first user. The first image can then be fed to one or more machine learning models in order to learn and capture the personal style of the first user. Subsequently, some embodiments capture the first user in another image or set of images. Some embodiments can then detect one or more features of the first user in these other images and based on the determining of the user's personal style in the first image, can apply data indicative of the personal style of the first user to the one or more features of the user in these other images.

Abstract: A method of ultrasound image volume reconstruction includes: providing a convolutional neural network (“CNN”); receiving a first dataset comprising at least one pair of consecutive ultrasound images; inputting the first dataset to the CNN; training the CNN with the first dataset; receiving a second dataset comprising an ultrasound video comprising a plurality of consecutive ultrasound images; inputting the second dataset to the CNN; and processing, by the CNN, the second dataset to produce as output a reconstructed 3D ultrasound image volume.

Abstract: According to one embodiment, an information processing device for a gesture-based ordering taking system or the like includes a processor. The processor is configured to detect a person making a particular gesture in image data from a camera, identify each person making the particular gesture in the image data as a first type person, then identify each first type person meeting a candidate condition in the image data as a candidate. The processor then designates a candidate as a subject according to a subject selection condition, and then detects gestured-based input from the subject in additional image data from the camera after the subject has been designated.

Abstract: The disclosure relates to methods and systems for determining one or more parameters, attributes, or characteristics in a meat sample using an iteratively trained detection model. In one embodiment, the methods and systems disclosed herein use automated methods to determine muscle fascicle fracturing. Methods and systems disclosed herein comprise processing one or more images of a meat sample captured with a data capture device; and using an iteratively trained detection model to determine an objective classification, including the presence or absence of muscle fascicle fracturing in the images of the meat sample.

Abstract: This disclosure relates generally to assessing soil carbon sequestration. One of the driving factors for climate change and global warming is emission of greenhouse gas emissions. Of the possible ways to reduce climate change and global warming adoption, sustainable agricultural practices will enable efficient soil carbon sequestration thereby reducing the greenhouse gases as well as increasing the crop yield. The disclosure is a method and a system for real time assessing soil carbon sequestration of farm based on remote sensing. The soil carbon sequestration of farm is assessed by continuously monitoring the farm at real time based on remote sensing using a plurality of satellite data and a plurality of farming data using several techniques. The several techniques utilized for assessing soil carbon sequestration includes machine learning, a carbon sequestration estimation technique, estimating a crop health index and an adoption index and computing a set of carbon sequestration parameters.

Abstract: The video based detection of pulse waveform includes systems, devices, methods, and computer-readable instructions for capturing a video stream including a sequence of frames, processing each frame of the video stream to spatially locate a region of interest, cropping each frame of the video stream to encapsulate the region of interest, processing the sequence of frames, by a 3-dimensional convolutional neural network, to determine the spatial and temporal dimensions of each frame of the sequence of frames and to produce a pulse waveform point for each frame of the sequence of frames, and generating a time series of pulse waveform points to generate the pulse waveform of the subject for the sequence of frames.

Abstract: Disclosed are a method and apparatus for training a neural network model to increase performance of the neural network model, the method including receiving input data and target data, pooling, by a neural network model, on a feature map extracted from the input data based on a probability for each of classes of the feature map, generating output data by inputting the input data to a neural network model, determining a loss based on comparing the output data and the target data and an auxiliary loss of the pooling, and training the neural network model based on the loss.

Abstract: [Problem] To provide a motion condition estimation device, a motion condition estimation method and a motion condition estimation program capable of accurately estimating the motion condition of monitored subjects even in a crowded environment. [Solution] A motion condition estimation device according to the present invention is provided with a quantity estimating means and a motion condition estimating means. The quantity estimating means uses a plurality of chronologically consecutive images to estimate a quantity of monitored subjects for each local region in each image. The motion condition estimating means estimates the motion condition of the monitored subjects from chronological changes in the quantities estimated in each local region.

Abstract: A method of classification of a film non-uniformity on a substrate includes obtaining a color image of a substrate with the color image comprising a plurality of color channels, obtaining a standard color for the color image of the substrate, for each respective pixel along a path in the color image determining a difference vector between the a color of the respective pixel and the standard color to generate a sequence of difference vectors, and sorting the pixels along the path into a plurality of regions including at least one normal region and at least one abnormal region based on the sequence of difference vectors, including comparing a multiplicity of the difference vectors in the sequence to a threshold.

Abstract: The image analysis device includes an analysis unit that analyzes input image data using a preset analysis function, and a control unit that, in a case in which a start-up request is received from an exterior, determines an analysis function to be a start-up target among plural analysis functions based on an analysis result obtained by the preset analysis function and a preset condition, and performs control to start-up the determined analysis function.

Abstract: In an embodiment, an intelligent system includes an electronic circuit configured to execute a neural network, to detect at least one feature in an image of a body portion while executing the neural network, and to determine a respective position and a respective class of each of the detected at least one feature while executing the neural network. For example, such a system can execute a neural network to detect at least one feature in an image of a lung, to determine a respective position within the image of each detected feature, and to classify each of the detected features as one of the following: A-line, B-line, pleural line, consolidation, and pleural effusion.

Abstract: A method of X-ray digital image correlation is provided. The method comprises aligning cameras of an X-ray imaging system, wherein the X-ray imaging system comprises two X-ray sources pointed at a target area at a stereo angle and two corresponding X-ray detectors behind the target area. The X-ray imaging system is tuned to determine power levels of the X-ray sources that maximize image contrast and signal-to-noise ratio. The system is calibrated to determine intrinsic and extrinsic stereoscopic imaging parameters. A specimen with a random contrast pattern comprising an X-ray attenuating material is placed in the target area. Stereoscopic X-ray images are taken of the specimen and processed according to the power levels of the X-ray sources to maximize image contrast and signal-to-noise ratio and remove background objects. Kinematic quantities are determined according to changes in the contrast pattern over a number of successive images.