Real‐time recognition and tracking for augmented reality books

Cho, Kyusung; Jung, Jaeho; Lee, Sangwook; Lim, Sang Ok; Yang, Hyun Seung

doi:10.1002/cav.431

Cited by 11 publications

(6 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Markerbased approach is the most commonly used registration method in the majority of applied AR systems. The markers are designed for easily detected and the shape is very regular, such as rectangular and circle [Cho et al 2011]. Kyusung Cho proposes an AR system for the interesting book reading.…”

Section: Related Workmentioning

confidence: 99%

Matching optimization based on KLT algorithm in natural feature tracking of augmented reality

Zhang

Yue

et al. 2012

Proceedings of the 11th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and Its Applications in Industry

View full text Add to dashboard Cite

Registration is one of the most difficult problems in augmented reality systems. One optimized method is proposed in this paper to remove the mismatched points. This method is based on the KLT tracking and uses the Fast algorithm to extract the natural feature. Math verification is given based on the small changes between the continuous frames captured by camera. An AR Reading system is achieved in this paper and SURF algorithm is exploited as page recognition process. The initial camera pose could be calculated in page recognition process. Experiments show that the optimization presented could improve the robustness of the registration and decrease the re-projection error.

show abstract

Section: Related Workmentioning

confidence: 99%

Matching optimization based on KLT algorithm in natural feature tracking of augmented reality

Zhang

Yue

et al. 2012

Proceedings of the 11th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and Its Applications in Industry

View full text Add to dashboard Cite

show abstract

“…Efforts to increase the number of recognition targets were first made on a personal computer (PC). In [15], Cho et al proposed the generic random forest (GRF) to implement an AR book with up to 200 pages on a PC. The GRF maximizes the reusability of the existing RF so that both object recognition and keypoint matching can be performed simultaneously.…”

Section: Related Workmentioning

confidence: 99%

Real-time Object Recognition with Pose Initialization for Large-scale Standalone Mobile Augmented Reality

Lee¹

2020

KSII TIIS

View full text Add to dashboard Cite

Mobile devices such as smartphones are very attractive targets for augmented reality (AR) services, but their limited resources make it difficult to increase the number of objects to be recognized. When the recognition process is scaled to a large number of objects, it typically requires significant computation time and memory. Therefore, most large-scale mobile AR systems rely on a server to outsource recognition process to a high-performance PC, but this limits the scenarios available in the AR services. As a part of realizing large-scale standalone mobile AR, this paper presents a solution to the problem of accuracy, memory, and speed for large-scale object recognition. To this end, we design our own basic feature and realize spatial locality, selective feature extraction, rough pose estimation, and selective feature matching. Experiments are performed to verify the appropriateness of the proposed method for realizing large-scale standalone mobile AR in terms of efficiency and accuracy.

show abstract

“…LGRF achieves these goals by combining two recently proposed techniques, generic randomised trees (GRT) [4] and lightweight random ferns (LRF) [5], which are extensions of RT and RF, respectively.…”

Section: Lightweight Generic Random Fernsmentioning

confidence: 99%

“…To implement multi‐target AR on mobile devices, both object recognition and keypoint matching have to be performed in real time under low memory constraints. LGRF achieves these goals by combining two recently proposed techniques, generic randomised trees (GRT) [4] and lightweight random ferns (LRF) [5], which are extensions of RT and RF, respectively.…”

Section: Lightweight Generic Random Fernsmentioning

confidence: 99%

Lightweight generic random ferns for multi‐target augmented reality on mobile devices

Lee

Yang

2013

Electron. lett.

View full text Add to dashboard Cite

Proposed use lightweight generic random ferns (LGRF), a fast keypoint classifier designed for multi-target augmented reality (AR) on mobile devices. LGRF uses binary features of image patches for both object recognition and keypoint matching of multiple objects, and stores probabilities in a single bit representation to reduce memory requirements. As a result, LGRF can perform simultaneous object recognition and keypoint matching in real time with low memory consumption, making it suitable for multi-target AR on mobile devices.Introduction: A core challenge in augmented reality (AR) is the efficient calculation of positions for virtual objects on real-world images. In this regard, keypoints have proven to be more useful than other image features, because of their robustness under viewpoint variation and partial occlusion. As keypoint recognition has become more commonplace, two major approaches have emerged.The first approach relies on local descriptors that are meant to be invariant, or at least robust, over affine deformations. Of these, the SIFT descriptor proposed in [1] has performed remarkably well and become a benchmark for other techniques. Unfortunately, like these other techniques, SIFT incurs a high computational cost at runtime, and is not yet suitable for real-time applications.The second approach recasts keypoint recognition as a classification problem. In short, the set of all possible appearances of each keypoint is considered a class, and is used to train a classifier. By shifting the computational burden from runtime to an offline training phase, this approach can achieve real-time keypoint recognition. In [2, 3], randomised trees (RT) and random ferns (RF) proved fast enough for frame-rate performance at high rates of recognition, making them suitable for AR applications. Unfortunately, these methods require significant memory at runtime (all the more when applied to multiple targets), and are further limited by their serialisation of object recognition and keypoint matching. For these reasons, they are ill suited to current mobile hardware.In this Letter, we propose a new extension of RF called lightweight generic random ferns (LGRF). This method directly addresses the memory and performance problems of RT/RF, and is very well suited to current mobile hardware.

show abstract

Real‐time recognition and tracking for augmented reality books

Cited by 11 publications

References 21 publications

Matching optimization based on KLT algorithm in natural feature tracking of augmented reality

Matching optimization based on KLT algorithm in natural feature tracking of augmented reality

Real-time Object Recognition with Pose Initialization for Large-scale Standalone Mobile Augmented Reality

Lightweight generic random ferns for multi‐target augmented reality on mobile devices

Contact Info

Product

Resources

About