Abstract:Irregular boundaries in image stitching naturally occur due to freely moving cameras. To deal with this problem, existing methods focus on optimizing mesh warping to make boundaries regular using the traditional explicit solution. However, previous methods always depend on hand-crafted features (e.g., keypoints and line segments). Thus, failures often happen in overlapping regions without distinctive features. In this paper, we address this problem by proposing RecStitchNet, a reasonable and effective network for image stitching with rectangular boundaries. Considering that both stitching and imposing rectangularity are non-trivial tasks in the learning-based framework, we propose a three-step progressive learning based strategy, which not only simplifies this task, but gradually achieves a good balance between stitching and imposing rectangularity. In the first step, we perform initial stitching by a pre-trained state-of-the-art image stitching model, to produce initially warped stitching results without considering the boundary constraint. Then, we use a regression network with a comprehensive objective regarding mesh, perception, and shape to further encourage the stitched meshes to have rectangular boundaries with high content fidelity. Finally, we propose an unsupervised instance-wise optimization strategy to refine the stitched meshes iteratively, which can effectively improve the stitching results in terms of feature alignment, as well as boundary and structure preservation. Due to the lack of stitching datasets and the difficulty of label generation, we propose to generate a stitching dataset with rectangular stitched images as pseudo-ground-truth labels, and the performance upper bound induced from the it can be broken by our unsupervised refinement. Qualitative and quantitative results and evaluations demonstrate the advantages of our method over the state-of-the-art.

Multilayered Stitch Generating for Random-Needle Embroidery

StitchGeneration: Modeling and Creation of Random-Needle Embroidery Based on Markov Chain Model

Stitches Generation for Random-Needle Embroidery Based on Markov Chain Model.

Paint with Stitches: a Style Definition and Image-Based Rendering Method for Random-Needle Embroidery

Vector and Pixel Space Depended Stitch Definition and Style Transfer for Random-Needle Embroidery

Stitch-Based Image Stylization for Thread Art Using Sparse Modeling.

Image-based Irregular Needling Embroidery Rendering

A Constrained Multi-sample Texture Synthesis Algorithm Based on Random Expanding of Circular Patches

Research on Approach of Generating Random Stitch in Embroidery CAD

A Controllable Stitch Layout Strategy for Random Needle Embroidery

Remote Sensing Image Stitch Using Modified Structure Deformation

Image Stylization for Thread Art via Color Quantization and Sparse Modeling.

FSIS: Fast and Seamless Image Stitching of High-Resolution Dunhuang Murals

RecStitchNet: Learning to Stitch Images with Rectangular Boundaries

Stitching Patch Arrangement Planning for Simulation of Irregular Needling Embroidery

Automatic Embroidery Texture Synthesis for Garment Design and Online Display

Towards End-to-End Embroidery Style Generation: A Paired Dataset and Benchmark

A Novel Method for Weft and Warp Yarn Segmentation in Multicolour Yarn‐dyed Fabric Images

Generating Embroidery Patterns Using Image-to-Image Translation

Digital 3D Smocking Design

Shape Decomposition Algorithm in Embroidery