Gesture Identification Using 2D Imaging Size Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current gesture identification methods using two-dimensional image collection units suffer from low accuracy due to the high cost of high-accuracy depth cameras and the inability to effectively distinguish between operations made within and outside a defined operating region.
Innovation Solution
A method that processes two-dimensional images to determine the imaging size of a pointing object, allowing for the identification of target images within a predetermined size range, thereby excluding operations made outside the operating region and improving accuracy by determining and executing instructions based on these target images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a high-accuracy depth camera is used to track the target and identify gestures, then the gesture identification accuracy is improved, but the device cost increases significantly
Solution Approach 1:
The patent uses a 2D image collection unit to capture images of the pointing object, then processes these 2D images to extract depth information and identify gestures. This approach copies the functionality of expensive depth cameras by using affordable 2D cameras combined with image processing algorithms, thereby achieving gesture recognition without requiring costly depth sensing hardware
Solution Approach 2:
The patent replaces the mechanical/optical depth sensing system (depth camera) with an computational imaging approach. Instead of using specialized depth-sensing hardware, the system uses 2D image processing algorithms to infer depth information and gesture patterns, substituting physical measurement mechanisms with computational methods
2Ease of manufacture
If a common two-dimensional image collection unit is used to identify gestures, then the device cost is reduced, but the gesture identification accuracy deteriorates
Solution Approach 1:
The patent changes the processing parameters of 2D images by analyzing imaging size variations of the pointing object across multiple 2D images. By extracting depth information from size changes in 2D images and processing these parameters through algorithms, the system achieves accurate gesture identification using only inexpensive 2D image collection units
Solution Approach 2:
The patent extracts depth information (third dimension) from 2D images by analyzing the imaging size of the pointing object. This allows the system to obtain 3D gesture data from 2D images, effectively adding depth dimensionality through computational methods rather than specialized hardware
3Reliability
If all 2D images are processed for gesture identification, then comprehensive gesture detection is achieved, but the processing time and computational resources increase
Solution Approach 1:
The patent extracts only the relevant 2D images that contain the pointing object from the sequence of 2D images, rather than processing all images. By identifying and extracting only the target images for gesture analysis, the system reduces computational workload and processing time while maintaining comprehensive gesture detection capability
Solution Approach 2:
The patent processes only a subset of 2D images that are relevant to gesture identification (those containing the pointing object within the operating region), rather than processing the entire image sequence. This partial processing approach reduces time loss while still achieving reliable gesture detection
Data Source
AI summary
A gesture identification method and an electronic device are provided. The gesture identification method includes: processing each 2D image of a plurality of 2D images including a pointing object, to acquire an imaging size of the pointing object in each 2D image; determining at least one target 2D image from the plurality of 2D images based on the imaging size of the pointing object in each 2D image; and determining and executing a corresponding instruction based on the at least one target 2D image.