Tri-map Generation for Image Matting Using Confidence Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image matting techniques struggle to accurately separate foreground from background, especially for objects like hair strands and fur, and require user intervention or multiple inputs, leading to complex and error-prone matte generation processes.
Innovation Solution
An image processing system detects objects using object detection techniques and generates a tri-map based on confidence maps computed from pre-trained deep learning models, depth sensors, or Time of Flight sensors, determining transformation values and labeling pixels as foreground, background, or unknown without human intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing image matting techniques are used, then foreground and background can be separated, but boundary detail accuracy deteriorates for objects like hair strands and fur
Solution Approach 1:
The patent segments the image processing task into multiple stages: object detection, confidence map generation, tri-map creation, and alpha matte computation. This segmentation allows each stage to focus on specific aspects (e.g., boundary regions in tri-map) to improve overall boundary detail accuracy for complex objects like hair and fur
Solution Approach 2:
The patent introduces intermediate structures (confidence map and tri-map) that mediate between the input image and final alpha matte. The tri-map specifically acts as an intermediary that identifies boundary regions, guiding the alpha matte computation to pay special attention to hair strands and fur details
2Measurement precision
If separate tri-map input or user intervention is used, then unknown pixel regions can be identified, but process complexity increases
Solution Approach 1:
The system performs self-service by automatically generating the tri-map from the input image through object detection and confidence map computation, eliminating the need for external tri-map input or user intervention. The algorithm autonomously identifies boundary regions and unknown pixels
Solution Approach 2:
The patent performs preliminary actions by pre-computing object detection results and confidence maps before the main alpha matte generation process. This preliminary processing automatically prepares the tri-map data structure, reducing the need for user intervention during the main processing stage
3Ease of operation
If automatic tri-map generation is implemented, then user intervention is reduced, but manufacturing precision of alpha matte deteriorates
Solution Approach 1:
The system uses feedback mechanisms where confidence maps (derived from object detection) inform the tri-map generation, which in turn guides the alpha matte computation. This feedback loop ensures that automatic generation maintains high precision by continuously refining results based on detected object boundaries and confidence levels
Solution Approach 2:
The patent employs parameter changes by adjusting transformation values based on image characteristics and object properties. These parameter adjustments optimize the automatic tri-map generation process to maintain alpha matte quality while requiring no user intervention
Data Source
AI summary
A system and a method of performing an image matting on an image are provided. The method includes detecting, by an image processing system, one or more objects in the image; determining, by the image processing system, a confidence map associated with the image using one or more image segmentation techniques for each of the one or more objects; and generating, by the image processing system, a tri-map for each of the one or more objects in the image from the confidence map based on at least one of a size of each of the one or more objects in the image and a distance between a first pixel in the image and a second pixel in at least one of the one or more objects in the image, wherein the tri-map is used to perform the image matting.