A scene matching algorithm based on SURF feature

Su, Juan; Qingsong, Xu; Jinghua, Zhu

doi:10.1109/iasp.2010.5476080

Cited by 10 publications

(4 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A major group is "keypoint"-based descriptors [10]. SIFT and speeded-up robust features (SURF) [11], [12] belong to the most commonly used keypoint detectors and feature descriptors for various applications [13], [14]. While most keypoint detection schemes extract real-valued feature vectors, algorithms such as binary robust invariant scalable keypoints (BRISK) [15] use binary feature descriptors for which image search queries need to be performed within a short time [16].…”

Section: A Image Descriptorsmentioning

confidence: 99%

Representing Medical Images With Encoded Local Projections

Tizhoosh

Babaie²

2018

IEEE Trans. Biomed. Eng.

View full text Add to dashboard Cite

This paper introduces the "encoded local projections" (ELP) as a new dense-sampling image descriptor for search and classification problems. The gradient changes of multiple projections in local windows of gray-level images are encoded to build a histogram that captures spatial projection patterns. Using projections is a conventional technique in both medical imaging and computer vision. Furthermore, powerful dense-sampling methods, such as local binary patterns and the histogram of oriented gradients, are widely used for image classification and recognition. Inspired by many achievements of such existing descriptors, we explore the design of a new class of histogram-based descriptors with particular applications in medical imaging. We experiment with three public datasets (IRMA, Kimia Path24, and CT Emphysema) to comparatively evaluate the performance of ELP histograms. In light of the tremendous success of deep architectures, we also compare the results with deep features generated by pretrained networks. The results are quite encouraging as the ELP descriptor can surpass both conventional and deep descriptors in performance in several experimental settings.

show abstract

Section: A Image Descriptorsmentioning

confidence: 99%

Representing Medical Images With Encoded Local Projections

Tizhoosh

Babaie²

2018

IEEE Trans. Biomed. Eng.

View full text Add to dashboard Cite

show abstract

“…Taking into account the high efficiency and the complexity of the matching algorithm, the method based on the nearest distance in the two directions [13] is adopted. As shown in formula (4) Based on the parameter derived from training, if the conditions shown in formula (4)(5)(6) are satisfied, the target is considered to be in the image .…”

Section: The Progressive Local Recognition Algorithmmentioning

confidence: 99%

Implementation of a progressive target recognition system for all the angles

Zhou

2012

International Conference on Automatic Control and Artificial Intelligence (ACAI 2012)

View full text Add to dashboard Cite

3D target recognition is a popular research field in the laser near-field detection, where the scan image of the target is obtained line by line. The arbitrary spatial position between the detector and the target, along with the interference caused by the clouds, make it difficult to realize accurate real-time identification for the detector. This paper presents a method to generate gray-scale plane images under any viewpoint by calculating the echo signals from the surface of 3D planes. Besides, considering both scan imaging and signal processing in hardware, a local feature based on improved and simplified SURF algorithm is implemented in hardware to describe the target. Then a progressive 3D target recognition system is built on the FPGA, and the experiments performed on the Modelsim simulation software show that the system can reach a high recognition rate.

show abstract

“…Su Juan et al [2] used SURF algorithm without preprocessing; the result of which was not satisfactory when initial position changed. Shugao Ma et al [3] applied local feature extraction and used Earth Mover Distance (EMD) for scene conformation which is not a robust method in remote sensing applications.…”

Section: Introductionmentioning

confidence: 98%

A Novel Scene Matching Method in Very High Resolution Satellite Images

Abolfazli

Moomivand

2011

2011 Seventh International Conference on Signal Image Technology &Amp; Internet-Based Systems

View full text Add to dashboard Cite

Scene matching which plays an important role in machine vision and remote sensing can be obtained from very high resolution satellite images. Although scale invariant feature transform (SIFT) is a powerful tool in scene conformation under different image status, but it is not adequate to give away a robust matching, due to the pettiness nature of objects in satellite images. In this paper, the scene matching algorithm has been modified using a combination of SIFT and bilateral filter along with a novel matching method. In this method, a bilateral filter is initially imposed on image and then the SIFT keypoints are extracted from the main and query image and finally imposing the proposed novel match method on the image resulted in acquiring a robust matching. Testing the method on a set of satellite images showed a vast decrease in false matching points and an enormous increase in processing speed which best satisfied the real-time requirements in comparison with results, using SIFT alone.

show abstract

A scene matching algorithm based on SURF feature

Cited by 10 publications

References 5 publications

Representing Medical Images With Encoded Local Projections

Representing Medical Images With Encoded Local Projections

Implementation of a progressive target recognition system for all the angles

A Novel Scene Matching Method in Very High Resolution Satellite Images

Contact Info

Product

Resources

About