Gesture Detection Using Face-Frame Filtering for False Hand Rejection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing gesture detection systems inaccurately identify objects other than hands as hands due to variations in video quality, leading to incorrect hand detection.
Innovation Solution
A gesture detection apparatus that includes a face frame information acquisition unit, a hand candidate detection unit, and a determination unit to set a face frame around the face and reject hand candidates based on overlap conditions with the face frame, ensuring accurate hand detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If gesture detection is performed based on video capture without additional constraints, then the detection coverage is wide, but the detection accuracy deteriorates due to false identification of non-hand objects
Solution Approach 1:
The detection process is segmented into multiple independent stages: face detection and framing, hand candidate detection, and overlap-based filtering. By dividing the detection task into separate modules with distinct functions, the system improves hand detection accuracy while maintaining manageable complexity through modular architecture.
Solution Approach 2:
A face frame is introduced as an intermediary element between the video input and hand detection output. The face frame serves as a reference structure that mediates the filtering of hand candidates through overlap detection, enabling accurate hand identification without requiring complex direct analysis of hand features alone.
2Area of stationary object
If the detection area is expanded to cover more potential hand regions, then the detection coverage increases, but the number of false positives increases
Solution Approach 1:
The system applies different detection strategies to different spatial regions: the face frame is detected with high precision in its specific location, while hand candidates are screened based on their spatial relationship (overlap) with the face frame. This localized quality approach ensures that the expanded detection area does not compromise reliability, as each region is processed with appropriate scrutiny.
3Productivity
If simple hand candidate detection is used without verification, then the processing speed is fast, but the detection accuracy is low
Solution Approach 1:
The system performs preliminary detection of hand candidates quickly across the video frame, then applies a verification step by checking overlap with the pre-detected face frame. This preliminary action approach maintains fast initial processing speed while adding a lightweight verification layer that significantly improves detection accuracy without requiring complex reprocessing.
Data Source
AI summary
Provided is a gesture detection apparatus that accurately detects a hand of an occupant in gesturing. The gesture detection apparatus includes a face frame information acquisition unit, a hand candidate detection unit, and a determination unit. The face frame information acquisition unit acquires face frame information. The face frame is set so as to surround the face of the occupant detected on the basis of the video. The hand candidate detection unit detects a hand candidate on the basis of the video. The determination unit rejects the information of the hand candidate so that the hand candidate is not detected as the hand of the occupant in the gesture of the occupant as a detection target on the basis of a predetermined condition regarding the overlap between the face frame in the video and the hand candidate frame set to surround the hand candidate.