Circular Padding for Seamless AIGC Panoramic Image Generation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current AIGC image generation technologies introduce splicing seams when processing 2D panoramic expansion views, leading to poor image quality and a compromised user experience when restoring 360° panoramic images.

Innovation Solution

The proposed method involves using an AIGC image generation model with a circular padding method for convolutional layers, which allows for seamless splicing by treating the image as cylindrical, eliminating vertical seams and improving image continuity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If conventional padding methods are used in AIGC image generation models, then the model can process panoramic images, but vertical splicing seams appear in the generated images

Engineering Contradiction:

Improveimage qualityVSAvoidsplicing seams

Core Design Contradiction:

Manufacturing precisionVSObject-generated harmful factors

Solution Approach 1:

The patent applies circular padding instead of conventional padding methods. In circular padding, pixels from one edge of the image are wrapped around to the opposite edge, creating a continuous cyclic structure. This inversion of the conventional padding approach eliminates the discontinuity that causes vertical splicing seams in panoramic image generation.

Inventive Principle:

Principle #13The other way round (Inversion)

Solution Approach 2:

The patent models the panoramic image as a cylindrical surface where the left and right edges are connected. By applying circular padding, the flat 2D image processing is transformed into curved surface processing, allowing seamless splicing at the edges and eliminating vertical seams in the generated panoramic images.

Inventive Principle:

Principle #14Spheroidality (Curvature)

2Ease of operation

If standard convolutional layer padding is used, then the AIGC model can generate panoramic images, but image continuity is compromised at spliced regions

Engineering Contradiction:

Improveimage generation capabilityVSAvoidimage continuity

Core Design Contradiction:

Ease of operationVSStability of the object's composition

Solution Approach 1:

Instead of using standard padding that creates discontinuities at image edges, the patent employs circular padding that wraps edge pixels to the opposite side. This inverted approach maintains image continuity across spliced regions while preserving the full functionality of the AIGC model for generating panoramic images.

Inventive Principle:

Principle #13The other way round (Inversion)

Solution Approach 2:

The patent applies circular padding as a preliminary processing step before feeding images into the AIGC model. This pre-processing ensures that the input images already have continuous edge transitions, preventing seam formation during the generation process and maintaining image stability throughout the workflow.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250191122A1Image generation method for eliminating splicing seams, computer device and storage medium

Publication Date: 2025.06.12 ARASHI VISION INC

AI summary

Embodiments of the present application disclose an image generation method for eliminating a splicing seam. The method may include inputting input information into an AIGC image generation model; and obtaining a panoramic expansion view generated by the AIGC image generation model based on the input information, wherein a padding method of a convolutional layer in the AIGC image generation model comprises circular padding.