Opportunity
The commercialization and widespread adoption of autostereoscopic 3D displays face significant bottlenecks due to the complexity and cost associated with traditional multi-view image generation methods. Conventionally, creating multi-view images for 3D effects requires capturing a scene using an array of cameras, each oriented along different optical axes. This approach involves cumbersome setup, synchronization of optical parameters (e.g., zoom, focus) across cameras, and substantial storage and distribution challenges for multi-channel video data. As a result, there is a scarcity of 3D content, limiting the market potential for autostereoscopic monitors, 3D digital photo frames, and related products. The existing problem is the lack of a simple, efficient, and accessible method to generate multi-view images without relying on complex hardware setups like camera arrays. This patent addresses the need for a streamlined solution that can convert readily available 2D images into multi-view formats, thereby enabling broader 3D content creation and enhancing user experiences on autostereoscopic displays without requiring specialized equipment or extensive resources.
Technology
This patent introduces an innovative method to generate multiple multi-view images from a single 2D source image, simulating the effect of capturing a scene from different perspective directions as with a camera array. The core technology involves processing the source image’s pixels to create at least two multi-view images, each representing a distinct viewing angle for 3D perception. The method begins by acquiring a 2D source image composed of multiple pixels. It then automatically generates the multi-view images solely from these pixels, using a disparity estimation process to calculate offsets or disparities for each pixel based on weighted color components (e.g., red, green, blue) or equivalent representations like luminance and chrominance. These disparities are applied to shift pixels horizontally, creating the illusion of depth and perspective variation. To enhance visual quality, the initial disparity map may be filtered using low-pass filters (e.g., Hamming, Gaussian) to smooth abrupt changes and ensure continuity in the 3D effect. Additionally, a saliency estimator can be employed to identify relevant pixels (e.g., via edge detection) for processing, reducing computational load by focusing on areas that contribute most to the 3D perception. The generated multi-view images can be integrated into a single composite image using a mask function tailored for autostereoscopic displays, allowing for direct rendering and viewing without the need for glasses. The technology can be implemented as software running on computing units, hardware processing circuits (e.g., FPGAs), or embedded in devices, making it versatile for applications in image processing, display systems, and content creation.
Advantages
- Eliminates the need for complex camera arrays, reducing hardware costs and setup efforts.
- Simplifies storage and distribution by converting single 2D images into multi-view formats, easing 3D content creation.
- Enhances visual quality through disparity filtering and saliency estimation, ensuring smooth and realistic 3D effects.
- Compatible with autostereoscopic displays, enabling glasses-free 3D viewing experiences.
- Flexible implementation options, including software, hardware chips (e.g., FPGAs), and integration into existing systems.
- Adjustable parameters (e.g., disparity weights, filters) allow customization based on user preferences for 3D intensity.
- Supports real-time processing for video sequences, extending applicability to dynamic content.
- Reduces computational load by selectively processing significant pixels, improving efficiency.
Applications
- Autostereoscopic 3D displays for monitors, TVs, and digital photo frames.
- Conversion of existing 2D image and video libraries into 3D formats for entertainment and media.
- Integration into cameras and smartphones for real-time 3D image generation.
- Use in medical imaging, gaming, and virtual reality to enhance depth perception.
- Implementation in advertising and digital signage for engaging 3D visual content.
- Educational tools and simulations requiring multi-view perspectives.
- Hardware chips (e.g., FPGAs) for embedded systems in consumer electronics.
- Software applications for photo editing and 3D content creation by professionals and hobbyists.
