China 
Africa 
Explore chapters and articles related to this topic
Computer Vision for Microstructural Image Representation: Methods and Applications
Published in Jeffrey P. Simmons, Lawrence F. Drummy, Charles A. Bouman, Marc De Graef, Statistical Methods for Materials Science, 2019
Brian L. DeCost, Elizabeth A. Holm
In addition to the affine region detectors [683] based on the Harris corner detector [411], several other region detectors have been applied. The Difference of Gaussians (DoG) [591, 592] and related detectors locate blob-like interest points by searching for local extrema in various approximations of the multiscale image Laplacian. Blob-based interest point detectors often find complementary sets of interest points to corner-based detectors, so they are sometimes used in tandem [[1192]]. The maximally stable extremal regions (MSER) [652] detects distinctive regions by (effectively) adaptively thresholding the input image and searching for local regions that have a stable segmentation over a range of threshold values.
Perception
Published in Hanky Sjafrie, Introduction to Self-Driving Vehicle Technology, 2019
Maximally Stable Extremal Regions (MSER) [42] is a blob detection method that works by detecting the property changes of a region (or a set of connected pixels) relative to its surroundings. As shown in Figure 3.17(a) and Figure 3.17(b), MSER describes an image as a set of regions that are maximally stable, or virtually unchanged, despite intensity changes. In other words, it seeks to find regions that remain visible across a wide range of brightnesses. The MSER region is usually described using an ellipsoid that is fitted to the actual shape. Compared to SIFT, MSER is faster and invariant to affine transformations, such as skewing [51].
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).
Image feature detection using an improved implementation of maximally stable extremal regions for augmented reality applications
Published in International Journal of Image and Data Fusion, 2018
Neetika Gupta, Mukesh Kumar Rohil
An affine invariant feature detector widely used in computer vision applications is the Maximally Stable Extremal Regions (MSER) (Matas et al. 2004). It works successfully with comparatively fewer regions of interests per image and is therefore well applied for tasks like image retrieval at a larger scale (Nister and Stewenius 2006). MSER regions are highly stable regions detected by thresholding the image at different grey levels and determining the change in area with respect to the change in intensity of the connected components. The normalised area of the connected component with respect to the change of area determines the stability canon of the respective region. The applications of MSER have evolved in tasks like object recognition (Matas et al. 2004, Obdrzalek and Matas 2002), tracking (Donoser and Bischof 2006b) and has been extended to work upon colour (Forssen 2007) and volumetric images (Donoser and Bischof 2006a) for object recognition and object-tracking applications. However, MSER regions are not fully scale invariant (Yu and Morel 2011). Scale change in an image is not just considered as a homothetic transformation; it involves change of blur and subsampling of an image. Scale normalisation in MSER is based on the area of the detected extremal regions; as a result, the detector is not able to perform well with the drastic changes of the level line geometry due to blur. Therefore, implementation of MSER provides a potential scope of improvement in order to make the detector fully invariant to affine transformations.
An automated liver tumour segmentation and classification model by deep learning based approaches
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
Sayan Saha Roy, Shraban Roy, Prithwijit Mukherjee, Anisha Halder Roy
MSER is a method for detecting blobs in images. The MSER algorithm has been used to detect liver tumours in our study. From an image, the MSER algorithm extracts several covariant regions. It is a connected module of some grey-level image sets that are stable. It is intended to pick virtually identical regions using a diverse set of accesses. All pixels below a certain threshold are considered white, whereas all pixels above or equal to that threshold are considered black. Depending on what is specified, MSER can extract bright patches on dark backgrounds (Ramesh et al. 2017).