Depth Image Hand Gesture Recognition Without Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hand gesture recognition methods require special equipment and calibration, limiting convenience and accuracy, especially when recognizing hand shapes using color images, which have limited features due to single colors.
Innovation Solution
An apparatus and method for hand gesture recognition based on depth images, using a depth image acquiring unit, depth point classifying unit, and hand model matching unit, which classify depth points into hand portions and match a three-dimensional hand model using machine learning techniques like support vector machines and neural networks, without the need for calibration, allowing real-time recognition of finger motions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a special glove device is worn for hand gesture recognition, then recognition accuracy is improved, but user convenience deteriorates due to movement restrictions and calibration requirements
Solution Approach 1:
The patent creates a virtual three-dimensional hand model that copies and represents the user's actual hand geometry. Instead of requiring physical sensors on the hand, the system captures depth image data and generates a digital twin of the hand that can be manipulated and analyzed computationally, eliminating the need for special gloves while maintaining recognition accuracy
Solution Approach 2:
The patent replaces the mechanical sensor-based system (gloves with sensors) with a vision-based depth imaging system. The mechanical contact sensors are substituted by optical depth cameras that capture hand geometry non-invasively, and machine learning algorithms that process the visual data to recognize gestures without physical constraints on the user
2Device complexity
If color images are used for hand shape recognition, then device complexity is reduced, but feature extraction capability deteriorates due to single color limitation
Solution Approach 1:
The patent transitions from two-dimensional color images to three-dimensional depth images. By adding the depth dimension (Z-axis) to the traditional XY plane imaging, the system captures spatial geometry information that is completely absent in color images. This dimensional enhancement provides rich structural features for hand shape recognition while maintaining relative system simplicity
Solution Approach 2:
The patent changes the fundamental parameter being measured from color intensity (RGB values) to spatial distance (depth values). Instead of analyzing variations in light reflection colors, the system measures the actual geometric distances from the camera to different points on the hand, providing fundamentally different and more informative data for shape recognition
Data Source
AI summary
Disclosed is an apparatus for hand gesture recognition based on a depth image, which includes a depth image acquiring unit configured to acquire a depth image including a hand region, a depth point classifying unit configured to classify depth points of a hand region in the depth image according to a corresponding hand portion by means of a machine studying method, and a hand model matching unit configured to match a three-dimensional hand model with the classified depth points by using distances between the classified depth points and a hand portion respectively corresponding to the depth points. A recognition method using the apparatus is also disclosed.