Digital Image Shadow Editing With Lighting-Separated ML Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital content modification systems struggle with accuracy and flexibility in editing complex shadows, particularly in portrait images, often altering both subject and background lighting conditions and lacking precise control over shadow intensity, shape, and position.
Innovation Solution
A shadow editing system utilizing deep learning techniques, including a lighting estimation network, lighting diffusion network, and shadow synthesis network, separates shadows from lighting representation, allowing users to modify dominant light parameters such as intensity, size, and position, and synthesizes adaptable shadows while preserving ambient lighting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional digital content modification systems are used to edit shadows, then the editing process is simple, but the accuracy and flexibility in editing complex shadows is poor
Solution Approach 1:
The patent segments shadow editing into distinct controllable parameters: intensity, shape, and position. This allows independent adjustment of each parameter through dedicated control interfaces, enabling precise shadow editing without requiring complex manual manipulation of pixel data. The segmentation of shadow attributes into separate control dimensions directly addresses the accuracy-f Flexibility contradiction by providing granular control over shadow properties.
Solution Approach 2:
The system implements parameter-based shadow control by allowing users to modify specific shadow parameters (intensity, shape, position) independently. This parameter change approach enables flexible shadow editing by adjusting numerical values that directly control shadow characteristics, rather than requiring complex graphical manipulation. The parameter change principle resolves the contradiction by providing simple numerical controls that yield precise shadow editing results.
2Ease of operation
If conventional systems modify images, then the process is straightforward, but they lack precise control over shadow intensity, shape, and position
Solution Approach 1:
The patent implements dynamic shadow editing by allowing real-time adjustment of shadow parameters through interactive controls. The system dynamically updates shadow intensity, shape, and position based on user input, enabling flexible experimentation with different shadow configurations. This dynamic approach maintains ease of operation while achieving precise shadow parameter control, as users can iteratively adjust parameters and immediately see results.
Solution Approach 2:
The system incorporates feedback mechanisms that display shadow preview results as parameters are adjusted. This feedback loop allows users to see the effect of parameter changes in real-time, enabling precise control over shadow characteristics while maintaining intuitive operation. The feedback principle resolves the contradiction by providing visual confirmation that links simple parameter adjustments to precise shadow editing outcomes.
3Adaptability or versatility
If conventional systems edit shadows, then the process is simple, but they alter both subject and background lighting conditions
Solution Approach 1:
The patent extracts shadow editing from global lighting modification by implementing selective shadow parameter control. The system separates shadow adjustments from overall lighting changes, allowing users to modify shadow characteristics without affecting ambient lighting conditions. This extraction principle enables adaptable lighting control by isolating shadow parameters from general lighting parameters, resolving the contradiction between versatility and complexity.
Solution Approach 2:
The system applies local quality control by allowing shadow parameter adjustments to affect only specific regions (subject areas) rather than the entire image. Users can control shadow intensity, shape, and position locally on subjects while preserving background lighting conditions. This local quality approach provides versatile lighting control without requiring complex global lighting modifications, as changes are confined to targeted areas.
Data Source
AI summary
The present disclosure relates to systems, non-transitory computer-readable media, and methods for editing shadows in digital images. In particular, in some embodiments, the disclosed systems determine, utilizing a lighting estimation network, an environment map for a digital image, the environment map comprising a dominant light. In addition, in some embodiments, the disclosed systems generate, utilizing a lighting diffusion network, a diffused image from the digital image, the diffused image comprising smoothed shading. Moreover, in some embodiments, the disclosed systems generate, utilizing a shadow synthesis network, a shadowed image from the diffused image and a modified environment map comprising a modified dominant light. Furthermore, in some embodiments, the disclosed systems generate, from the diffused image and the shadowed image, a modified digital image comprising an edited shadow.