Image Blur Correction Device Panning Assist Face Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image blur correction technologies face challenges in accurately detecting motion vectors during panning, particularly for novice photographers, leading to potential over-correction and subject shake, especially when photographing moving subjects like people, due to irregular motions of hands and feet, which result in low reliability motion vectors.
Innovation Solution
An image blur correction device with a calculation unit that calculates subject motion using shake detection signals and face detection positions, employing a panning assist mode to differentiate between subject and background vectors, and a control unit that adjusts shutter speed based on angular velocity data to correct image blur, ensuring reliable panning assistance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If a long shutter speed is set to obtain an impression of dynamism in the subject, then the sense of dynamism increases because an amount of background flow increases, but camera shake or subject shake easily occurs
Solution Approach 1:
The system performs preliminary face detection and motion vector calculation before the actual photographing to determine appropriate shutter speed settings. By detecting the subject's face position and calculating motion vectors in advance, the system can pre-calculate the necessary correction values and shutter speed parameters, allowing the camera to be ready for optimal dynamic imaging without causing shake during the actual capture.
2Measurement precision
If motion vectors of the entire body of a person including hands and feet are used, then the angular velocity of a subject to be calculated may be erroneously calculated, but using only face region provides more reliable motion data
Solution Approach 1:
The system divides the subject detection into distinct segments: face detection and body motion detection. By focusing on the face region specifically, the system excludes irregular motions of hands and feet from the calculation, thereby improving angular velocity measurement precision. The face detection unit and motion vector calculation unit work together to segment the analysis, considering only the face area for motion vector computation while maintaining overall subject tracking.
3Productivity
If block matching is used to detect motion vectors, then motion detection can be performed, but low contrast conditions result in detection errors
Solution Approach 1:
The system applies different detection strategies to different regions of the image. Instead of uniformly applying block matching across the entire image, the system focuses on the face region where motion vectors are most reliable. The face detection unit identifies the face area, and the motion vector calculation unit then applies correlation operations specifically within this region, where contrast and motion characteristics are more favorable for accurate detection.
Data Source
AI summary
An imaging device detects a shake detection signal and an amount of motion of an image within a photography screen when a panning assist mode is set. A search range of a face detection position of a subject is set, and a process of changing a calculation process of a subject vector is performed based on a face detection result. If the face detection position of the subject is acquired, a first calculation process in which a subject vector is calculated from motion vectors within a first detection range based on the face detection position is performed. If the face detection position of the subject is not acquired and a subject vector can be detected, a second calculation process n which a subject vector is calculated from the motion vectors within a second detection range based on a focus detection frame set within the photography screen is performed.