Image Analysis TRIZ Cases

A single-lens 3-D imaging device uses a central aperture and off-axis defocusing apertures to capture spatially separated image samples on one sensor.

Parallel raster scans generate optical flow estimates, reducing processing and memory resources.

Transmission-mode x-ray imaging acquires coherent diffraction patterns from volume defects within three-dimensional structures.

A display control unit arranges virtual objects near real objects using template scene graphs to define spatial relationships.

A method segments video frames to transmit only essential visual portions, reducing data volume.

A computing device analyzes color, infrared, and depth data streams to differentiate authentic faces from forged representations.

An object tracking system leverages terminal location history to detect relevant cameras for video analysis.

A display apparatus segments storage regions for each indicator to enable independent virtual object pasting operations.

An event detection device generates signals from pixel changes to analyze facial motion sequences.

Weight painting drives automatic UV map partitioning that scales texture islands by gradient scores, eliminating manual coordinate manipulation.

A vehicle driving assist apparatus adjusts attention display timing based on predicted reaching time and lateral distance.

Machine learning predicts multi-dimensional substrate geometry and stochastic edge placement errors from aerial images.

A video analysis system calculates user dance matching rates against reference videos.

Periodic decoding at key frames combined with correlation filter tracking reduces computational load and energy consumption while maintaining high precision.

Transfer system moves imaging data between upstream and downstream detection units to adjust parameters without halting production lines.

A merged reality system streams color and depth video from multiple vantage points to enable dynamic rendering of virtual and real-world objects.

A hybrid light measurement method combines laser stripe and fringe pattern projection using MEMS technology.

A subject region extraction unit calculates pixel value differences between infrared images to separate foreground objects from backgrounds.

Automatic outline tracing generates two-dimensional polygons from digital images to create interactive entertainment assets without further user input.

Selective depth helper data reduces visually disturbing errors by transmitting partial depth maps only where local generation quality falls below a threshold.

Dynamic dimension compression of partial feature quantities reduces storage requirements and power consumption while maintaining recognition accuracy.

Artificial vision system detects service fabric and adhesive tape residues on tyre semi-finished products using brightness threshold analysis.

Deep learning object recognition drives camera view control to maintain surveillance accuracy when targets move out of the field of view.

A system preloads video with closest viewpoint information using Media Presentation Description data.

Merging adjacent convex polygons reduces storage requirements and computation time while preserving edge fidelity in maskless lithography patterning.

A client device anchors augmented reality objects to physical surfaces using precise three-dimensional coordinate detection.

A holographic diffractive optical element generates a dot pattern that a multiple beam grating splits into equal intensity replicas.

A 3D display device composites images using a depth suitability determiner to align additional content with main visuals.

Convolutional neural networks identify drivable path geometries directly, eliminating post-processing steps that increase computing power and reduce efficiency.

Distance-specific stereo camera parameters compensate for water refraction to enable accurate fish size measurement.

A vehicle parking display control system generates segmented guidance lines superimposed on rear camera video to support intuitive vehicle handling.

A camera and inertial sensor fusion system determines treated body portions using pictorial representations and acceleration data.

A mobile rendering system creates synthetic images from weighted object mesh subsets to visualize arbitrary viewpoints.

A training data selection device filters sample images using feature amount comparison to store only verified appropriate data.

A gaze tracking device determines parameters by processing image data into reference and viewing angle vectors.

A computing system calculates perceptibility scores for virtual universe content using cube map visibility analysis.

Adjustable circuitry manages virtual boundaries between digital ink input and video output regions on a single display panel.

Mixed-scale tokenization reduces computational complexity and memory demands by assigning optimal resolution scales to specific image regions.

A recurrent coherent network processes dynamic vision sensor frames to generate candidate detection boxes.

Segmentation and arm removal correct patient table absorption to resolve head-to-toe diameter variations for accurate dose estimation.

An interactive map interface links active sensors to target entities through direct user selection and confirmation actions.

A dynamic expression model refines tracking parameters from RGB-D data to generate real-time facial animation.

A hemispherical clamping system segments image-based light data to exclude back-side regions from sampling.

A microparticle sorting device uses an image sensor and controller to adjust driving voltage for stable droplet streams.

A device selects frame images based on calculated rotation angles to generate moving image data with consistent rotational speed.

A saliency map creation method decomposes images into frequential sub-bands and estimates movement to simulate human visual processing.

ANAGDA filter fuses inertial and terrain data to reduce error accumulation in GPS-denied environments.

A medical article recognition system extracts local feature vectors from video frames to identify devices and pharmaceuticals in real time.

Merge correlated candidates into composite objects to resolve classification accuracy versus system complexity trade-offs.