Face Detection Activation via Pixel Depth Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Face detection operations in electronic devices, such as smartphones, consume excessive computation and power due to unnecessary processing, as they aimlessly perform face detection on every image.
Innovation Solution
A face detection activating method that uses depth values from a subset of pixels within the pixel array to determine whether to activate the face detection operation, focusing on specific depth ranges and concentration of pixels to minimize unnecessary computation and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face detection operation is performed on every image, then face detection accuracy is maintained, but power consumption and unnecessary computation increase significantly
Solution Approach 1:
The patent applies preliminary action by performing depth estimation on selected pixels before face detection. The system first estimates depth information from a subset of pixels, then uses this depth data to determine whether face detection should be activated. This preliminary depth analysis allows the system to avoid unnecessary full face detection operations while maintaining detection accuracy when faces are present.
Solution Approach 2:
The patent implements partial action by using only a subset of pixels for depth estimation rather than processing all pixels in the image. Specifically, depth information is obtained from selected pixels (e.g., regularly spaced pixels or pixels in specific regions), which provides sufficient depth data for activation decisions while significantly reducing computational load and power consumption compared to full-image processing.
2Reliability
If face detection operation is performed on every image, then detection coverage is maximized, but computation time and processing load increase
Solution Approach 1:
The system performs preliminary depth estimation before face detection to quickly assess whether face detection is warranted. This preliminary step uses computationally efficient depth calculations on selected pixels, allowing the system to filter out images that definitely do not contain faces before committing to the more computationally intensive face detection process.
Solution Approach 2:
The patent introduces depth information as an intermediary between image capture and face detection. Instead of directly performing face detection on every image, the system first extracts depth information from selected pixels and uses this intermediate data to make activation decisions. This intermediary step enables intelligent filtering while maintaining detection coverage.
3Measurement precision
If depth values from all pixels are used to determine activation, then activation accuracy is improved, but power consumption and processing time increase
Solution Approach 1:
The patent applies partial action by using depth values from only a subset of pixels rather than all pixels in the image array. The system selects specific pixels (such as regularly spaced pixels or pixels in representative regions) to estimate depth information, which provides sufficient accuracy for determining whether to activate face detection while significantly reducing the computational power required compared to using all pixels.
Solution Approach 2:
The patent implements segmentation by dividing the pixel array into multiple regions or groups and selecting representative pixels from each segment for depth estimation. This segmented approach allows the system to maintain spatial coverage and activation accuracy while processing far fewer pixels than would be required to analyze the entire image, thereby reducing power consumption.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present application provides a face detection activating method, applied in an electronic device comprising a pixel array. The face detection activating method comprises computing a plurality of first depth values corresponding to a plurality of first pixels within the pixel array, wherein the plurality of first pixels is a part of pixels in the pixel array; and determining whether Lo activate a face detection operation according to the plurality of first depth values.