Image Processing Subject Region Expansion via Dual Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle to accurately trim images by removing unnecessary backgrounds and objects while ensuring the intended subject is not cut off, especially when the subject is posing in a non-standard manner or holding objects.
Innovation Solution
An image processing apparatus that employs two detection methods: a first method for initial subject detection and a second method, such as salient region detection or binarization, to update the subject region, ensuring the subject is not cut off and minimizing inclusion of unnecessary elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the subject region is estimated only based on object detection results, then the trimming process is simple and fast, but the intended subject is cut off when the subject is posing in non-standard manner
Solution Approach 1:
The patent applies preliminary action by performing object detection first to obtain an initial subject region, then using salient region detection as a preliminary step before final trimming. This preliminary salient region detection identifies areas that should be included in the subject region, allowing the system to correct inaccuracies from object detection alone while maintaining efficient processing speeds.
Solution Approach 2:
The patent implements feedback by using the salient region detection results to feedback-correct the object detection results. The salient region information is fed back into the trimming process to adjust and expand the subject region when necessary, ensuring that the intended subject is not cut off while maintaining processing efficiency.
2Reliability
If a uniform margin is added to the object detection result to prevent subject truncation, then the subject is protected from being cut off, but unnecessary margins and unnecessary objects are included
Solution Approach 1:
The patent applies local quality by differentiating the treatment of different regions around the detected subject. Instead of adding a uniform margin to the entire object detection result, the system selectively expands the subject region only in areas where salient region detection indicates the subject may be truncated. This localized expansion prevents subject truncation while avoiding the inclusion of unnecessary margins and objects in areas where the subject is clearly defined.
3Manufacturing precision
If multiple detection methods are combined to improve subject region accuracy, then the subject region accuracy is improved, but the processing time and complexity increase
Solution Approach 1:
The patent applies segmentation by dividing the detection process into distinct functional segments: object detection for initial subject identification, salient region detection for accuracy improvement, and trimming for final processing. Each segment handles a specific aspect of the problem, allowing the system to achieve high accuracy by combining the strengths of different detection methods while managing processing time through efficient segmentation of tasks.
Data Source
AI summary
An image processing apparatus includes an input unit that inputs an image, and a processor configured to read out a program stored in a memory, and executes the program. The processor is configured to detect an intended subject from the input image by a first detection method, set an intended subject region for the detected intended subject, detect the intended subject from the input image by a second detection method different from the first detection method, and update the set intended subject region by using a detection result of the second detection method.