A neurophysiologically driven system segments images into chips and assigns target probabilities using EEG data.
Angular boundary pattern matching identifies real objects using extracted geometric features from image data.
Sensor-based PPE monitoring tracks usage limits to automate compliance alerts, preventing hazards from overlooked equipment disposal.
A pretrained machine learning system adapts to shifted data distributions using a non-saturating surrogate loss function.
A database-driven annotation system manages data facets to streamline AI training dataset preparation.
Visual and acoustic fingerprints trigger automatic gameplay recording from a buffer without manual player intervention.
A parking lot tracing controller generates and corrects line templates using similarity calculations to determine accurate positions.
Removes redundant filters via diversity analysis to lower battery consumption while maintaining classification accuracy.
A biometric virtual account system generates secure financial access using physiological measurements to replace physical gift cards.
Sparse coding and Laplacian eigenvectors create invariant representations for neural network image recognition.
A LiDAR object detection method segments point clouds using geometric features to determine representative division points for cluster separation.
A two-pass convolutional neural network generates high-resolution object proposals using forward and backward passes.
A mobile input method integrates barcode scanning and optical character recognition into a unified interface for direct data entry.
A computing device dynamically adjusts neural network configurations to classify physical features in specific deployment environments.
A lane line extraction system merges laser point cloud data with camera images to determine accurate road markings.
A computing system generates real-time performance scores by combining live judge observations with recorded review data.
Deploying a local recognition model on the terminal generates object labels autonomously, reducing server workload and protecting user privacy.
An email relay apparatus converts moving images into compatible formats to enable display on terminals with restricted capabilities.
A face swapping apparatus uses trained identity and attribute encoders to transform facial features into compatible vectors for image synthesis.
Reconstructs partial fingerprint images by extrapolating finger motion parameters, resolving edge accuracy issues caused by limited sensor size.
An optical sensor acquires point data from reflection points to determine fog density based on distance distributions.
A machine learning model predicts normal system behavior using only standard sensor data to identify deviations.
A multifaceted detection process enhances barcode recognition using image enhancement and auxiliary information processing.
A facial recognition sign-in system captures user face images to verify identity against stored registration data.
Automated image sharing identifies subjects via facial recognition and transmits copies using contact databases, eliminating manual distribution delays.