KAZE Feature Research Articles

3D point cloud reconstruction using panoramic images

Panorama photogrammetry, the process of analyzing panoramic images, has gained popularity in close-range photogrammetry for 3D reconstruction over the past decade. Initially, researchers utilized cylindrical or spherical panoramic images created through specialized cameras or conventional ones with rectilinear lenses. However, these methods were hindered by the high cost of panorama equipment and the need for manual reconstruction. Consequently, there's a growing demand for automated algorithms capable of reconstructing 3D point clouds from stitched panorama images. This study aims to provide a cost-effective solution for automatic 3D point cloud reconstruction from panoramas. The study is divided into two parts; it first outlines an image acquisition strategy for capturing overlapping perspective images to facilitate accurate panorama generation. Subsequently, it introduces an automated algorithm for 3D point cloud reconstruction from panorama images. The process utilizes the KAZE feature detector for feature detection and introduces a novel feature matching approach for matching panorama images. Accuracy assessment of the reconstructed 3D point clouds was done using three methods: Line Segment Based approach, yielding RMSE errors of 34.2mm and 39mm for dataset-1 and dataset-2 respectively, No-Reference 3D Point Cloud Quality Assessment, resulting in quality scores of 8.5939 and 7.4535 for dataset-1 and dataset-2 respectively, and M3C2 distance method computed value of 0.091059 and 0.165179 respectively, indicating high quality of the generated point clouds.

Palmprint recognition system using VR-LBP and KAZE features for better recognition accuracy

The palmprint recognition system has gained significant attention in security and law enforcement due to its unique features, such as principle lines, ridges, and wrinkles. However, many existing methods for extracting these features have limited accuracy, especially when the image illumination varies or the size of the processed pixels increases. Previous studies have shown that the local binary patterns (LBP) algorithm is effective for palmprint recognition due to the rich texture characteristics of a palmprint. In this paper, we propose a new technique for a robust contact-based palmprint identification system using vertical-LBP and KAZE feature detection. Our technique aims to improve recognition accuracy by using KAZE, which is a nonlinear diffusion approach that extracts nonlinear features from the evolution of the illuminance of an image. We also utilize principal component analysis (PCA) to reduce the dimensionality of the generated descriptor vector elements. The proposed method was tested on the PolyU database and achieved recognition accuracy of 99.7%.

Soft tissue monitoring of the surgical field: detection and tracking of breast surface deformations.

Deformable object tracking is common in the computer vision field, with applications typically focusing on nonrigid shape detection and usually not requiring specific 3D point localization.In surgical guidance however, accurate navigation is intrinsically linked to precise correspondence of tissue structure. This work presents a contactless, automated fiducial acquisition method using stereo video of the operating field to provide reliable fiducial localization for an image guidance framework in breast conserving surgery. On n=8 breasts from healthy volunteers, the breast surface was measured throughout the full range of arm motion in a supine mock-surgical position. Using hand-drawn inked fiducials, adaptive thresholding, and KAZE feature matching, precise three-dimensional fiducial locations were detected and tracked through tool interference, partial and complete marker occlusions, significant displacements and nonrigid shape distortions. Compared to digitization with a conventional optically tracked stylus, fiducials were automatically localized with 1.6 ± 0.5 mm accuracy and the two measurement methods did not significantly differ. The algorithm provided an average false discovery rate <0.1% with all cases' rates below 0.2%. On average, 85.6 ± 5.9% of visible fiducials were automatically detected and tracked, and 99.1 ± 1.1% of frames provided only true positive fiducial measurements, which indicates the algorithm achieves a data stream that can be used for reliable on-line registration. Tracking is robust to occlusions, displacements, and most shape distortions. This work-flow friendly data collection method provides highly accurate and precise three-dimensional surface data to drive an image guidance system for breast conserving surgery.

Label-free technology for traceable identification of single green pepper through features in UV fluorescent images

In the food supply chain, vegetable traceability holds significant importance. Conventional methods for traceability rely on tag-based systems such as barcodes, QR codes, or RFID tags. However, these methods face challenges when it comes to tracing single vegetables, such as green peppers between the grading facility and greenhouse harvesting. Unlike fruits like melons and watermelons, green peppers lack visible unique features on their surface that can be used to identify individual vegetables. Through our research, we have discovered that the fluorescence images of green peppers display a unique texture on the surface, which provides the possibility of identifying single green pepper. We proposed a single pepper traceable method that combines pepper images from the greenhouse and postharvest stage under UV light using imaging features. Our experiments aimed to evaluate the method's performance, including feature description, tracing success rate, performance change with storage, and changes with different length of the green peppers. The results showed that the KAZE feature was suitable for describing the surface feature of a green pepper under UV light, achieving a feature-matching performance of 81.3 % success rate in tracking individual peppers in each of the 15 packages, each with 10 peppers, using images from the day after harvest and greenhouse images. Furthermore, the method's performance could be affected by the storage time and length of the peppers. The proposed method could be a cost-effective, accurate, and label-free method to achieve single green pepper level traceability in smart agriculture.

A computer vision approach to identifying the manufacturer of posterior thoracolumbar instrumentation systems.

Knowledge of the manufacturer of the previously implanted pedicle screw systems prior to revision spinal surgery may facilitate faster and safer surgery. Often, this information is unavailable because patients are referred by other centers or because of missing information in the patients' records. Recently, machine learning and computer vision have gained wider use in clinical applications. The authors propose a computer vision approach to classify posterior thoracolumbar instrumentation systems. Lateral and anteroposterior (AP) radiographs obtained in patients undergoing posterior thoracolumbar pedicle screw implantation for any indication at the authors' institution (2015-2021) were obtained. DICOM images were cropped to include both the pedicle screws and rods. Images were labeled with the manufacturer according to the operative record. Multiple feature detection methods were tested (SURF, MESR, and Minimum Eigenvalues); however, the bag-of-visual-words technique with KAZE feature detection was ultimately used to construct a computer vision support vector machine (SVM) classifier for lateral, AP, and fused lateral and AP images. Accuracy was tested using an 80%/20% training/testing pseudorandom split over 100 iterations. Using a reader study, the authors compared the model performance with the current practice of surgeons and manufacturer representatives identifying spinal hardware by visual inspection. Among the three image types, 355 lateral, 379 AP, and 338 fused radiographs were obtained. The five pedicle screw implants included in this study were the Globus Medical Creo, Medtronic Solera, NuVasive Reline, Stryker Xia, and DePuy Expedium. When the two most common manufacturers used at the authors' institution were binarily classified (Globus Medical and Medtronic), the accuracy rates for lateral, AP, and fused images were 93.15% ± 4.06%, 88.98% ± 4.08%, and 91.08% ± 5.30%, respectively. Classification accuracy decreased by approximately 10% with each additional manufacturer added. The multilevel five-way classification accuracy rates for lateral, AP, and fused images were 64.27% ± 5.13%, 60.95% ± 5.52%, and 65.90% ± 5.14%, respectively. In the reader study, the model performed five-way classification on 100 test images with 79% accuracy in 14 seconds, compared with an average of 44% accuracy in 20 minutes for two surgeons and three manufacturer representatives. The authors developed a KAZE feature detector with an SVM classifier that successfully identified posterior thoracolumbar hardware at five-level classification. The model performed more accurately and efficiently than the method currently used in clinical practice. The relative computational simplicity of this model, from input to output, may facilitate future prospective studies in the clinical setting.

Region specific and subimage based neighbour gradient feature extraction for robust periocular recognition

This paper proposes a Periocular recognition algorithm that uses region-specific and sub-image-based neighbor gradient feature extraction to achieve better recognition results. This approach initially segments the periocular region into four sub-regions such as eyebrow, eye corner regions, upper and lower eye fold regions after detecting the left and right eye corner points. KAZE feature extraction algorithm is used to extract the features from the upper eye fold region, while the HOG feature extraction algorithm is used to extract the feature from the eyebrow and eye corner regions. This approach also estimates the shape of the eyebrow by estimating the distance from N points on the eyebrow region to the eye corner midpoint. The eyebrow shape feature also contains the width and height measures through N points on the eyebrow that give its shape. The proposed approach also proposes a sub-image-based neighbor gradient (SING) feature extraction that extracts the neighbor gradient features from a 3×3 sub-image. Finally, the extracted features are trained using the Naïve Bayes classifier. The experimental evaluation was done using the AR dataset, CASIA Iris distance dataset, and UBIPr dataset using the metrics such as rank-1, rank-5 recognition accuracies, the area under the ROC curve (AUC), and equal error rate (EER). The proposed scheme provides a rank-1 recognition rate of 92.32 %, 97.41 %, and 97.87 % for the dataset UBIPr, CASIA-Iris, and AR datasets respectively. The experimental results reveal that the proposed method outperforms the traditional periocular recognition algorithms.

Quantitative analysis of visual codewords of a protein distance matrix.

3D protein structures can be analyzed using a distance matrix calculated as the pairwise distance between all Cα atoms in the protein model. Although researchers have efficiently used distance matrices to classify proteins and find homologous proteins, much less work has been done on quantitative analysis of distance matrix features. Therefore, the distance matrix was analyzed as gray scale image using KAZE feature extractor algorithm with Bag of Visual Words model. In this study, each protein was represented as a histogram of visual codewords. The analysis showed that a very small number of codewords (~1%) have a high relative frequency (> 0.25) and that the majority of codewords have a relative frequency around 0.05. We have also shown that there is a relationship between the frequency of codewords and the position of the features in a distance matrix. The codewords that are more frequent are located closer to the main diagonal. Less frequent codewords, on the other hand, are located in the corners of the distance matrix, far from the main diagonal. Moreover, the analysis showed a correlation between the number of unique codewords and the 3D repeats in the protein structure. The solenoid and tandem repeats proteins have a significantly lower number of unique codewords than the globular proteins. Finally, the codeword histograms and Support Vector Machine (SVM) classifier were used to classify solenoid and globular proteins. The result showed that the SVM classifier fed with codeword histograms correctly classified 352 out of 354 proteins.

Fault Detection of Permanent Magnet Synchronous Motor Based on Image Content Retrieval

The motor fault detection based on data-driven approach has achieved some results. However, there are still some problems, such as complex signal processing, long training time and poor adaptability to different working conditions. A new fault detection method for permanent magnet synchronous motor based on image content retrieval is proposed in this paper, which does not require complex signal processing and model training, and has strong adaptability to conditions. First, the fault signals are converted into images by the symmetrized dot pattern method. Second, an image coding technology based on KAZE feature and bag of features is proposed to reduce the computational cost. Then, to improve adaptability of conditions, a cyclic update algorithm of image retrieval library based on image fusion is proposed. Finally, fault detection is realized through image retrieval of feature code matching. Extensive experiments demonstrate the superiority and effectiveness of the proposed method.

A Novel Approach for Small Object Detection in Medical Images through Deep Ensemble Convolution Neural Network

Small objects detection in medical image becomes an interesting field of research that helps the medical practitioners to focus on in-depth evaluation of diseases. The accurate localization and classification of objects face tremendous difficulty due to lower intensity of the images and distraction of pixel points that vary the decision on identifying the shape, structure etc. In many real-time cases, detection and classification of tiny objects in the medically treated images becomes mandatory. The proposed system is designed in the same criteria in which the semantic segmentation of tiny objects in the medical images is considered. The system design focused on implementing the model for different kinds of human organs such as lung and liver. The axial CT or PET images of Lung and Liver are considered as the prime input for the given system. Detection of tiny objects in the CT-PET images, segmenting it from the background and classification of segmented part as Tumor or Nodule is discussed. The preprocessed images are feature extracted after the morphology segmentation that determines the structural features of the tiny object being segmented. The feature vectors are nothing but the feature points from Kaze feature extraction and Morphology segmented image. These two inputs are fetched to the Deep ensemble Convolution neural network (DECNN) to obtain the dual classification results. Performing the quantitative measurements to evaluate the decision making system for nodule or tumor class is determined. The performance measure is done using accuracy, precision, recall and F1Score.

Detection of the 3D temperature characteristics of maize under water stress using thermal and RGB-D cameras

Global warming and water resource shortage greatly influence the crop growth and negatively affect the crop yield. Breeding water–stress resistant crop varieties is one of the effective ways to handle this problem. The surface temperatures of the crop are essential for assessing their water stress resistance. Thermal imaging has been widely used to acquire the crop surface temperatures. However, most studies focus on 2D measurement. A 3D thermal imaging system is designed, and a method is developed by using the thermal and RGB-D data to acquire 3D thermal information of maize under water stress conditions. First, thermal and RGB-D cameras were used to collect the thermal and color images, and depth data of maize at the jointing stage and the thermal and color images were processed to extract the edge images of maize. Second, the KAZE feature was selected to register the edge images. Testing results showed that the KAZE feature has better performance than the SURF and BRISK features in the thermal and color image registration of maize. On the basis of the thermal and color image registration, the depth data of maize were assigned with temperature values. Third, the depth data of maize were further processed with denoising, maize extraction, amplification, smoothing, and temperature correction steps to improve the data qualities. Finally, the crop water stress index and canopy–air temperature difference values of each point were calculated. The results demonstrated that the system and the proposed method can effectively detect the water stress characteristics of maize in 3D, which can be combined with the morphological traits of leaves to synthetically analyze the water stress resistance of the crop.

Palmprint features matching based on KAZE feature detection

Palmprint is very popular biometric recognition system that is able to guarantee high accuracy. It has attracted increasing amount of attention because palmprints are abundant of many characteristics, such as the principle lines, ridges, minute points and textures for the use of images with low resolution. In this paper we propose palmprint feature detection based on KAZE technique. Palmprint texture has many important points for discrimination process. Selecting the best number of point using KAZE is very important for classification process in order to avoid overlapping features in different class. The experimental work has been done using polyU palmprint database in order to evaluate the best number of features.

A novel high precision mosaic method for sonar video sequence

The mosaic of sonar images is more difficult than the mosaic of traditional optical images due to their poor quality and the difficulty in extracting feature points. The existing mosaic methods of sonar images have a series of problems, such as low correct matching rate, large cumulative errors and high requirements for the quality of collected sonar images. In this paper, we proposed a high precision method to implement the mosaic of the underwater sonar video image sequence. Firstly, Accelerated Unsharp Masking (AUSM) algorithm is proposed to preprocess the original image. Then we extract KAZE feature points from preprocessed sonar images. A matching method combining multiple matching strategies with Progressive Sample Consensus (PROSAC) algorithm is followed to complete the image registration. Weighted fusion method and a region of interest (ROI) acquisition method based on the slope of right border is used to optimize the mosaicked image. Finally, we can obtain a high-quality panoramic image of underwater sonar video by a global mosaic strategy. Mosaic experiments on the sonar video image sequence collected by multi-beam sonar demonstrate that the proposed method in this paper can increase the number of feature points by about 70% and make the correct matching rate higher than 70%. The proposed method also has good robustness and the cumulative error during multi-image mosaic is less.

Automated hemangioma detection using Otsu based binarized Kaze features

This study aims to detect liver hemangioma on CT images by using hybrid image processing methods as well as binarized histogram of gradients based Kaze feature extraction. Using hybrid strategies, automatic hemangioma detector design is the state of art of this study. Our study helps doctors to detect the solid liver masses. Proposed algorithm includes detection of hemangioma using Otsu auto-threshold based Histogram of Gradients (HOG) and Kaze feature extraction implementation. 48 liver CT images, 28 of which are hemangiomas and 20 of which are healthy liver images, are used as the dataset. CT images are obtained by the Department of the Radiology at Firat University. Presented work was implemented to 48-CT images and 91,66% accuracy was achieved for different shaped and sized hemangiomas. These results show that the developed algorithm could ease the process of detecting liver masses for radiologist and doctors could evaluate their findings easily.

Applying large-scale PIV to water monitor discharge experiment

Industrial fires are undeniably the most dangerous type of fires for firefighters, and the master stream from water monitor is one of the safest means to deploy defensive or offensive strategies against such fires. However, few studies pay attention to the efficacy of the master stream. To address this issue, we extracted and analyzed numerical data from physical master streams. The proposed methods comprising of digital image processing (DIP) and large-scale particle image velocimetry (LS-PIV), which were conducted in the full-scale discharge in order to obtain coordinate trajectory pattern as well as the instantaneous and mean vector-velocity fields. The flow vectors in the different zones of the master stream were compared using KAZE feature detection. This study thus offers more extensive and detailed experimental data to validate computational fluid dynamics (CFD) simulations and opens an avenue for future fire-safety research.

KTRICT A KAZE Feature Extraction

To improve the retrieval accuracy in CBIR system means reducing this semantic gap. Reducing semantic is a necessity to build a better, trusted system, since CBIR systems are applied to a lot of fields that require utmost accuracy. Time constraint is also a very important factor since a fast CBIR system leads to a fast completion of different tasks. The aim of the paper is to build a CBIR system that provides high accuracy in lower time complexity and work towards bridging the aforementioned semantic gap. CBIR systems retrieve images that are related to query image (QI) from huge datasets. The traditional CBIR systems include two phases: feature extraction and similarity matching. Here, a technique called KTRICT, a KAZE-feature extraction, tree and random-projection indexing-based CBIR technique, is introduced which incorporates indexing after feature extraction. This reduces the retrieval time by a great extent and also saves memory. Indexing divides a search space into subspaces containing similar images together, thereby decreasing the overall retrieval time.

A computer vision approach to identifying the manufacturer and model of anterior cervical spinal hardware.

Recent advances in computer vision have revolutionized many aspects of society but have yet to find significant penetrance in neurosurgery. One proposed use for this technology is to aid in the identification of implanted spinal hardware. In revision operations, knowing the manufacturer and model of previously implanted fusion systems upfront can facilitate a faster and safer procedure, but this information is frequently unavailable or incomplete. The authors present one approach for the automated, high-accuracy classification of anterior cervical hardware fusion systems using computer vision. Patient records were searched for those who underwent anterior-posterior (AP) cervical radiography following anterior cervical discectomy and fusion (ACDF) at the authors' institution over a 10-year period (2008-2018). These images were then cropped and windowed to include just the cervical plating system. Images were then labeled with the appropriate manufacturer and system according to the operative record. A computer vision classifier was then constructed using the bag-of-visual-words technique and KAZE feature detection. Accuracy and validity were tested using an 80%/20% training/testing pseudorandom split over 100 iterations. A total of 321 total images were isolated containing 9 different ACDF systems from 5 different companies. The correct system was identified as the top choice in 91.5% ± 3.8% of the cases and one of the top 2 or 3 choices in 97.1% ± 2.0% and 98.4 ± 13% of the cases, respectively. Performance persisted despite the inclusion of variable sizes of hardware (i.e., 1-level, 2-level, and 3-level plates). Stratification by the size of hardware did not improve performance. A computer vision algorithm was trained to classify at least 9 different types of anterior cervical fusion systems using relatively sparse data sets and was demonstrated to perform with high accuracy. This represents one of many potential clinical applications of machine learning and computer vision in neurosurgical practice.

Partial Fingerprint Recognition of Feature Extraction and Improving Accelerated KAZE Feature Matching Algorithm

Nowadays for Personal Identification Numbers (PIN) and passwords an efficient and secure alternative has been used which is called e-security. Due to the various computer frauds like identity theft and computer hacking there is a rapid increase in the financial losses of payment. In present day, the banking sector is accepting the payments through credit cards for every on-line as well as offline transaction which actually encourages the cashless payments. It will be the foremost convenient because of do on-line looking out, bills payments etc. Therefore, there is also increase in the fraud risks that area associated with the misuse of the credit cards as a result of technology development. So we go for biometric security to resolve this problem. The aim of our study is to given security through biometric technology of fingerprint recognition. Minutiae extraction and minutiae matching technique of AKAZE to boost the accuracy and high speed computation.

Detection of copy-move forgery using AKAZE and SIFT keypoint extraction

Digital image manipulation techniques are becoming increasingly sophisticated and widespread. Copy-move forgery is one of the frequently used manipulation techniques. In this paper, we propose a keypoint based copy-move forgery detection (CMFD) technique, which is a combination of accelerated KAZE (AKAZE) and scale invariant feature transform (SIFT) features. By using AKZAE and SIFT, a significant number of keypoints are extracted even in a smooth region to detect the manipulated regions efficiently. After formation of the mixed keypoints, the g2NN is used for matching process to locate the duplicated regions. The experimental results show that the proposed method can detect the duplicated regions even if the image is post-processed with scaling, rotation, noise and JPEG compression operations. To validate the robustness and effectiveness of the proposed method, a statistical analysis is performed using the ANOVA method.

Automatic Dewarping of Retina Images in Adaptive Optics Confocal Scanning Laser Ophthalmoscope

Retina images acquired by an adaptive optics confocal scanning laser ophthalmoscope (AOSLO) usually need to remove image warp to improve image quality. The most significant task of AOSLO image dewarping is image registration. Most traditional feature-based registration algorithms used for AOSLO images are based on Gaussian scale space. However, the homogeneous Gaussian blurring reduces the localization accuracy of feature points and the distinctiveness of feature descriptors. In this paper, the accelerated KAZE (AKAZE) feature based on nonlinear scale space was utilized to register AOSLO retinal images for the first time, and an efficient strategy based on matched feature points for frame selection was proposed to automatically accomplish AOSLO retinal image dewarping. Moreover, a flexible method based on power spectra analysis is proposed to study the minimum number of frames needed to accomplish image dewarping. The extensive experiments demonstrated that the AKAZE method is more suitable for AOSLO image dewarping, benefitted with better accuracy, robustness, and rapidity compared with several traditional registration methods based on Gaussian scale space.

AKAZE-Based Visual Odometry From Floor Images Supported by Acceleration Models

To realize the self-localization of autonomous robots, methods for the 2D motion measurement of robots are required. In this research, a self-localization system using a CCD camera is proposed. In the proposed system, the self-localization is estimated by movement tracking using some keypoints detected from the floor images captured by the CCD camera. For the illumination of floor image, two LED illuminate are used. These lighting systems are installed in such a way that lit from both sides of the floor and parallel to the floor so that some minimal bumps or original veins are captured from the floor. An accelerated KAZE feature is applied in the proposed system for keypoint detection and for computing resulting in the generation of the descriptor. In the estimation, the location information from the previous step and the estimated position one frame ahead are used. Using the estimated result and computed descriptor, the proposed system matches the nearest keypoint between the estimated keypoint detected from the previous floor image and the keypoint detected from the next floor image at the beginning. The Hamming distance is employed in the proposed system to evaluate the matching. If the Hamming distance is longer than a threshold, the proposed system tries to match from the 2nd nearest keypoint. Based on an experiment in which the measurement distance and computation time are investigated, the effectiveness of the proposed method is confirmed.