Semi-Supervised Image Stitching from Unstructured Camera Arrays
Image stitching involves combining multiple images of the same scene captured from different viewpoints into a single image with an expanded field of view. While this technique has various applications in computer vision, traditional methods rely on the successive stitching of image pairs taken from...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-11-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/23/9481 |
| _version_ | 1827592031537463296 |
|---|---|
| author | Erman Nghonda Tchinda Maximillian Kealoha Panoff Danielle Tchuinkou Kwadjo Christophe Bobda |
| author_facet | Erman Nghonda Tchinda Maximillian Kealoha Panoff Danielle Tchuinkou Kwadjo Christophe Bobda |
| author_sort | Erman Nghonda Tchinda |
| collection | DOAJ |
| description | Image stitching involves combining multiple images of the same scene captured from different viewpoints into a single image with an expanded field of view. While this technique has various applications in computer vision, traditional methods rely on the successive stitching of image pairs taken from multiple cameras. While this approach is effective for organized camera arrays, it can pose challenges for unstructured ones, especially when handling scene overlaps. This paper presents a deep learning-based approach for stitching images from large unstructured camera sets covering complex scenes. Our method processes images concurrently by using the <i>SandFall</i> algorithm to transform data from multiple cameras into a reduced fixed array, thereby minimizing data loss. A customized convolutional neural network then processes these data to produce the final image. By stitching images simultaneously, our method avoids the potential cascading errors seen in sequential pairwise stitching while offering improved time efficiency. In addition, we detail an unsupervised training method for the network utilizing metrics from Generative Adversarial Networks supplemented with supervised learning. Our testing revealed that the proposed approach operates in roughly ∼1/7th the time of many traditional methods on both CPU and GPU platforms, achieving results consistent with established methods. |
| first_indexed | 2024-03-09T01:42:58Z |
| format | Article |
| id | doaj.art-57d2b2e6e2e14094a2813bbb80944697 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-09T01:42:58Z |
| publishDate | 2023-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-57d2b2e6e2e14094a2813bbb809446972023-12-08T15:26:09ZengMDPI AGSensors1424-82202023-11-012323948110.3390/s23239481Semi-Supervised Image Stitching from Unstructured Camera ArraysErman Nghonda Tchinda0Maximillian Kealoha Panoff1Danielle Tchuinkou Kwadjo2Christophe Bobda3Department of Electrical and Computer Engineering, The Herbert Wertheim College of Engineering, University of Florida, Gainesville, FL 32611-6200, USADepartment of Electrical and Computer Engineering, The Herbert Wertheim College of Engineering, University of Florida, Gainesville, FL 32611-6200, USADepartment of Electrical and Computer Engineering, The Herbert Wertheim College of Engineering, University of Florida, Gainesville, FL 32611-6200, USADepartment of Electrical and Computer Engineering, The Herbert Wertheim College of Engineering, University of Florida, Gainesville, FL 32611-6200, USAImage stitching involves combining multiple images of the same scene captured from different viewpoints into a single image with an expanded field of view. While this technique has various applications in computer vision, traditional methods rely on the successive stitching of image pairs taken from multiple cameras. While this approach is effective for organized camera arrays, it can pose challenges for unstructured ones, especially when handling scene overlaps. This paper presents a deep learning-based approach for stitching images from large unstructured camera sets covering complex scenes. Our method processes images concurrently by using the <i>SandFall</i> algorithm to transform data from multiple cameras into a reduced fixed array, thereby minimizing data loss. A customized convolutional neural network then processes these data to produce the final image. By stitching images simultaneously, our method avoids the potential cascading errors seen in sequential pairwise stitching while offering improved time efficiency. In addition, we detail an unsupervised training method for the network utilizing metrics from Generative Adversarial Networks supplemented with supervised learning. Our testing revealed that the proposed approach operates in roughly ∼1/7th the time of many traditional methods on both CPU and GPU platforms, achieving results consistent with established methods.https://www.mdpi.com/1424-8220/23/23/9481image stitchingself-supervised learningimage blendingunstructured camera arraysscene representation |
| spellingShingle | Erman Nghonda Tchinda Maximillian Kealoha Panoff Danielle Tchuinkou Kwadjo Christophe Bobda Semi-Supervised Image Stitching from Unstructured Camera Arrays Sensors image stitching self-supervised learning image blending unstructured camera arrays scene representation |
| title | Semi-Supervised Image Stitching from Unstructured Camera Arrays |
| title_full | Semi-Supervised Image Stitching from Unstructured Camera Arrays |
| title_fullStr | Semi-Supervised Image Stitching from Unstructured Camera Arrays |
| title_full_unstemmed | Semi-Supervised Image Stitching from Unstructured Camera Arrays |
| title_short | Semi-Supervised Image Stitching from Unstructured Camera Arrays |
| title_sort | semi supervised image stitching from unstructured camera arrays |
| topic | image stitching self-supervised learning image blending unstructured camera arrays scene representation |
| url | https://www.mdpi.com/1424-8220/23/23/9481 |
| work_keys_str_mv | AT ermannghondatchinda semisupervisedimagestitchingfromunstructuredcameraarrays AT maximilliankealohapanoff semisupervisedimagestitchingfromunstructuredcameraarrays AT danielletchuinkoukwadjo semisupervisedimagestitchingfromunstructuredcameraarrays AT christophebobda semisupervisedimagestitchingfromunstructuredcameraarrays |