China
Africa
Explore chapters and articles related to this topic
Automatic Screening of COVID-19 Based on CT Scan Images Through Extreme Gradient Boosting
Published in Mitul Kumar Ahirwal, Narendra D. Londhe, Anil Kumar, Artificial Intelligence Applications for Health Care, 2022
G. Madhu, A. Govardhan, K. Lakshman Srikanth, G. Nagachandrika
Herbert Bay et al. (2006) [43] presented the technique as a SURF feature descriptor that computes small portions of the allocated image and is based on feature detectors and multiscale space theory. This is determined by the Hessian Matrix factor, and the basic concept of SURF is to generate a scale-invariant local feature descriptor from input image data using any feature extraction method. The SURF algorithm [44] is like SIFT in that it has two major parts, the first of which uses squared cut streams and the Hessian matrix to find the image's focal points. Also, by obtaining constrained features about this stage, feature descriptors can be created. These are often generated by examining localized feature squared images and the region surrounding a point of interest (POI), and hence Haarwavelet responses within a specific neighborhood, as well as their answers at specific interval-based sampling points. SURF characteristics are pivot and scale-invariant but only have a small affine invariance. The mahotas SURF Python library is used in this study to compute local regions of blood images using feature extraction. The POI is calculated using the entire image as well as image regions that are very similar to the interest points that are being viewed for further study. Feature detection for a CT scan COVID-19 image is showed in Figure 8.6; major features are detected using the SURF technique, as shown in Figure 8.6.
Image Processing for Knowledge Management and Effective Information Extraction for Improved Cervical Cancer Diagnosis
Published in Kavita Taneja, Harmunish Taneja, Kuldeep Kumar, Arvind Selwal, Eng Lieh Ouh, Data Science and Innovations for Intelligent Systems, 2021
SURF is kind of feature detector and feature descriptor. This technique can be used by various applications such as object recognition and classification model. This is slightly similar to SIFT (Scale-Invariant Feature Transform). SURF is very advanced and fast version than SIFT. To detect the object in a image, it uses determinant of Hessian blob detector, which consists of three integer operator. SURF descriptor will be used to recognize the object, like face, car, anything and it is used to track the object using points of interest. SURF has major three phases:Interest point orientationNearest neighbor description componentsFeature matching
Vision-Based Tracking for a Human Following Mobile Robot
Published in Laxmidhar Behera, Swagat Kumar, Prem Kumar Patchaikani, Ranjith Ravindranathan Nair, Samrat Dutta, Intelligent Control of Robotic Systems, 2020
Laxmidhar Behera, Swagat Kumar, Prem Kumar Patchaikani, Ranjith Ravindranathan Nair, Samrat Dutta
A human-following robot needs to track a human walking in front of the robot using its on-board camera. The person may exhibit natural human motion including pose changes due to out-of-plane rotations. In this chapter, SURF is used as the visual feature, which is known to be robust to photometric and geometric distortions, and computationally efficient compared to other point features, like SIFT. The interest point based methods make use of object recognition for detecting a target in a given frame and are less influenced by abrupt object motion arising out of low frame rate or non-stationary camera. SURF-based tracking methods usually consist of two steps: (i) To represent the target or the reference model in terms of feature descriptors, and (ii) to infer the best location by computing the correspondences between the source frame and the target frame. The wrong correspondences are usually removed using RANSAC algorithm [470, 471].
A study of multi-target image-based displacement measurement approach for field testing of bridges
Published in Journal of Structural Integrity and Maintenance, 2022
Isaias A. Colombani, Bassem Andrawes
Like DIC, FBIR benefits from the arrival of high-resolution cameras, and perhaps even more so because of its ability to identify natural features in images, eliminating the need to apply patterns to a target object. FBIR utilizes feature-based algorithms for tracking deformations (Wang et al., 2015). An example of the feature identification algorithm used in this study is the popular SURF method presented by Bay et al. (2006) used for detecting distinct features in each image. The algorithm can identify features in an image that are invariant to scaling, translation, and rotation while being efficient computationally (Bay et al., 2006; Lowe, 1999). The M-estimator Sample Consensus (MSAC) algorithm was used to find the optimum set of matched points in the ROI by rejecting outliers in the dataset (Torr & Murray, 1997). The feature points can be filtered further by applying thresholds to exclude ambiguous matches, increase the feature point density, and improve the confidence of the feature matches. The set of identified feature points can then be averaged to obtain a displacement for the selected region.
Research status and prospect of visual image feature point detection in body measurements
Published in The Journal of The Textile Institute, 2022
Wenqian Feng, Yanli Hu, Xinrong Li, Yuzhuo Li
Comparing the three feature point detection algorithms described above, SIFT and SURF have scale-invariant features and rotation invariance, and their application scope is focused on image feature matching, such as visual Simultaneous Localization and Mapping (SLAM), 3D model establishment, and action recognition (Alwan & Ku-Mahamud, 2020; Hidalgo & Braunl, 2020; Xu et al., 2016). In terms of the accuracy and efficiency of feature point detection, SURF is more efficient and faster. Luo performed scale transformations on images and found that the offset of the feature point coordinates detected by the Harris algorithm was small, indicating better performance than SIFT (Luo, 2018). The Harris algorithm has rotation invariance, but is not scale-invariant, although it offers very good feature point detection (Vishwakarma & Bhuyan, 2020). To compare the feature point detection effects of the three algorithms on a human contour image, the algorithms were verified using OpenCV; the detection effects of SIFT and SURF algorithms are shown in Figure 2(a) and 2(b). The Harris corner detection algorithm is used to detect feature points. The number of feature points to be detected can be adjusted using a threshold value, with a smaller threshold resulting in more feature points. The detection effect when the threshold is set to 0.03 is shown in Figure 2(c).
A Training-Free Approach for Generic Object Detection
Published in IETE Journal of Research, 2022
Bhakti V. Baheti, Sanjay N. Talbar, Suhas S. Gajre
We would like to compare this LSS based feature extraction approach with some of the state-of-art sparse feature point extractors like Maximally Stable Extremal Regions(MSER) [25], Scale Invariant Feature Transform(SIFT) [10] and Speeded Up Robust Features(SURF) [11]. MSER is an algorithm for blob detection in images and is mostly used in object recognition and stereo matching. SIFT and SURF feature detectors compute distinctive invariant local features. SURF is basically an efficient alternative to SIFT in terms of speed and robustness. We implemented these feature extractors and results are shown in the figure below. Figure 6(a) shows a sample image of bottle.Figure 6(b)–(d) shows locations of extracted features with MSER, SIFT and SURF, respectively. These algorithms basically detect local features within image like blobs, corners or interest points. They do not as such contain information about self similarity of these features and hence the object shape. On the other hand, our proposed feature extraction scheme generates those feature locations containing strucrural information as shown in Figure 6(e).