Systems and method are claimed for forming an artifact attenuated seismic image. The method includes obtaining an input seismic image, selecting a seismic partition from the input seismic image and determining a seismic dip for the seismic partition. The method further includes determining flattened seismic partition from the seismic partition based, at least in part, on the seismic dip, determining a filtered seismic partition from the flattened seismic partition, and determining an unflattened seismic segment based on the filtered seismic partition. The method still further includes determining the artifact attenuated seismic image based on the unflattened seismic segment. The system includes a seismic source, a plurality of seismic receivers for detecting and recording an observed seismic dataset generated by the radiated seismic wave; and a seismic processor configured form the artifact attenuated seismic image.

A method for multi-target detection and an apparatus for multi-target detection are capable of detecting at least two targets in real time or near real time. The real-time detection or near real time detection can be achieved by at least one of a Recipe Group Approach, an End Member Grouping Approach, and a Pixelated Grouping Based Approach.

The present disclosure relates to systems, methods, and non-transitory computer readable media for generating shift-resilient neural network outputs based on utilizing a dense pooling layer, a low-pass filter layer, and a downsampling layer of a neural network. For example, the disclosed systems can generate a pooled feature map utilizing a dense pooling layer to densely pool feature values extracted from an input. The disclosed systems can further apply a low-pass filter to the pooled feature map to generate a shift-adaptive feature map. In addition, the disclosed systems can downsample the shift-adaptive feature map utilizing a downsampling layer. Based on the downsampled, shift-adaptive feature map, the disclosed systems can generate shift-resilient neural network outputs such as digital image classifications.

A system for determining occupancy in an environment is provided. The system includes plurality of sensor bundles, with each bundle including a presence sensor and a motion sensor. The system further includes a controller in communication with each sensor bundle. The controller is configured to designate one of the sensor bundles as presence triggered if persons are present within a field of view of the presence sensor. The controller is further configured to designate one of the sensor bundles as motion triggered if persons are moving within a field of view of the motion sensor. The controller is further configured to determine a triggered bundle count of the sensor bundles which are both presence triggered and motion triggered. The controller is further configured to determine an occupancy count for the environment based upon the triggered bundle count.

The Invention presents a system and a method for digitizing a human body shape from anthropometrical measurements. The proposed system and method allow reconstructing the 3D human body quickly and accurately, improving disadvantages of costly and timely traditional methods, which not only requires digitized persons to be naked or wear tight clothes but also could use hazardous lights to their health. The system in the invention includes two main modules and two supplementary blocks to reconstruct the 3D human body from anthropometric measurements, which are: (1) Input Block, (2) Pre-Processing Module, (3) Optimization Module, (4) Output Block. The method in the invention includes four steps: (1) Step 1a: collecting human body measurements, (2) Steps 1b: Initial Population; (3) Step 2: Optimizing; (4) Step 3: Displaying digitized human body shape.

The present invention relates to a scene descriptor generation system, including an image sensor including a plurality of pixels, successively supplying S binary image planes, each including a single binary value per pixel; and a processing circuit configured to, for each binary image plane supplied by the sensor (100), to implement the successive steps of: a) calculating N convolutions of the binary image plane with respectively N distinct convolution kernels, to provide N convolved images; b) generating at least one meta-image from the N convolved images supplied at step a); and c) for each of the N convolved images supplied at step a), generating a binary descriptor from said convolved image and from said at least one meta-image.

A method for predicting a clamp force using a convolutional neural network includes: generating a cepstrum image from a signal processing analysis apparatus; extracting a characteristic image by multiplying a predetermined weight value to pixels of the generated cepstrum image through artificial intelligence learning; extracting, as a representative image, the largest pixel from the extracted characteristic image; synthesizing an image by synthesizing the extracted representative image information; and predicting a clamp force by comparing the synthesized image with a predetermined value.

The device comprises at least one accelerometer, and a controller receiving input signals from the at least one accelerometer. The controller configured to filter stroke characteristics from the input signal using a filter module. The controller then applies a first statistical module on the filtered signal and obtains a first output signal. Due to the first statistical module, the first output signal is obtained, which is agnostic to type of swim stroke employed by the swimmer. The controller then determines the swim metric based on the first output signal and an adaptive threshold value. The swim metric is lap completion or lap count or turn event, during swimming by a swimmer. The device consumes less power and also agnostic to swim styles and turn styles employed by swimmers.

According to one aspect, the disclosure proposes a method for detecting events or features in physical signals by implementing an artificial neural network. The method includes evaluating the probability of presence of the event or feature by implementing the artificial neural network. The method includes implementing the artificial neural network in a nominal mode and to which a physical signal having a first so-called nominal resolution is fed, as long as the probability of the presence of the event or feature is below a threshold. The method further includes implementing the artificial neural network in a reduced consumption mode with a reduced resolution, as long as the probability of the presence of the event or feature is above the threshold. The reduced resolution is lower than the first resolution.

The present application provides a method, a device and a storage medium for predicting a remaining service life of a rail transit hardware device, the method including: generating particles of the hardware device at an initial moment; determining a state equation for a moment of prediction from a pre-established multi-stage state variation equation; determining particle weights at the moment of prediction on the basis of the state equation for the moment of prediction; predicting a state value at the moment of prediction according to the particle weights at the moment of prediction and the state equation for the moment of prediction; and determining the remaining service life of the hardware device on the basis of the state value at the moment of prediction.