2D feature detection via local energy

doi:10.1016/S0262-8856(96)01137-7

Image and Vision Computing

Volume 15, Issue 5, May 1997, Pages 353-368

https://doi.org/10.1016/S0262-8856(96)01137-7 Get rights and content

Abstract

Accurate detection and localisation of two-dimensional (2D) image features (or ‘key-points’) is important for vision tasks such as structure from motion, stereo matching and line labelling. Despite this interest, no one has produced an adequate definition of 2D image features that encompasses the variety of features that should be included under this banner. In this paper, we present a new method for the detection of 2D image features that relies upon maximal 2D order in the phase domain of the image signal. Points of maximal phase congruency correspond to all the different types of 2D features detected by other schemes, including grey-level corners, line terminations, and a variety of junctions. An assessment of our implementation's performance is provided, in terms of its robustness, accuracy of detection and localisation of 2D image features.

References (29)

G. Medioni et al.
Corner detection and curve representation using cubic b-splines
Computer Vision, Graphics and Image Processing
(1987)
L. Kitchen et al.
Gray-level corner detection, Pattern Recognition Letters
M.C. Morrone et al.
Feature detection from local energy
Pattern Recognition Letters
(1987)
S. Venkatesh et al.
On the classification of image features
Pattern Recognition Letters
(1990)
C.G. Harris
Determination of ego-motion from matched points
S.L. Hong et al.
Stereo correspondence: A hierarchical approach
L. Zusne
Visual Perception of Form
(1970)
J.F. Canny
Finding edges and lines in images
H. Asada et al.
The curvature primal sketch
IEEE Trans. PAMI
(1986)
P.R. Beaudet
Rotationally invariant image operators

W. Förstner

A framework for low level feature extraction

R. Deriche et al.

Accurate corner detection: an analytical study

R. Deriche et al.

A computational approach to corner and vertex detection

International Journal of Computer Vision

(1993)

K. Rohr

Modelling and identification of characteristic intensity variations

Image and Vision Computing

(February 1992)

Cited by (68)

PCNet: A structure similarity enhancement method for multispectral and multimodal image registration
2023, Information Fusion
Multispectral and multimodal images are of important usage in the field of multi-source visual information fusion. Due to the alternation or movement of image devices, the acquired multispectral and multimodal images are usually misaligned, and hence image registration is pre-requisite. Different from the registration of common images, the registration of multispectral or multimodal images is a challenging problem due to the nonlinear variation of intensity and gradient. To cope with this challenge, we propose the phase congruency network (PCNet) to enhance the structure similarity of multispectral or multimodal images. The images can then be aligned using the similarity-enhanced feature maps produced by the network. PCNet is constructed under the inspiration of the well-known phase congruency. The network embeds the phase congruency prior into two simple trainable layers and series of modified learnable Gabor kernels. Thanks to the prior knowledge, once trained, PCNet is applicable on a variety of multispectral and multimodal data such as flash/no-flash and RGB/NIR images without additional further tuning. The prior also makes the network lightweight. The trainable parameters of PCNet is 2400× and 1500×less than the deep-learning registration method deep homography network (DHN) and unsupervised deep homography network (UDHN), while its registration performance surpasses them. Experimental results validate that PCNet outperforms current state-of-the-art conventional multimodal registration algorithms. Besides, PCNet can act as a complementary part of the deep-learning registration methods, which significantly boosts their registration accuracy. On the Columbia imaging and vision laboratory (CAVE) multispectral dataset, the percentage of the number of images under 1 pixel average corner error (ACE) of UDHN is raised from 0.1% to 82.5% after the processing of PCNet.
A novel OCT servoing control method for endomicroscopy surface scanning
2023, Biomedical Signal Processing and Control
Probe-based confocal laser endomicroscopy (pCLE) enables cell-level visualization for in-vivo tissue of the organ surface in real time. However, an image sequence is needed to generate a large-scale mosaic because the field of view (FOV) of pCLE is restricted. Image mosaicking solves the small FOV issue but only for high-quality pCLE images. The key to acquiring high-quality images is maintaining a constant contact distance even if the surface of the tissue is significantly rough. In this work, optical coherence tomography (OCT) is used to perform surface scanning by maintaining constant contact between the probe tip and the tissue. We proposed a distance detection method to find the position of the tissue surface in OCT images and used a position-based OCT servoing control scheme to guide the surface scanning. Experimental results show that the proposed control scheme with OCT guidance can obtain a large and clear circular mosaic (the diameter is 1 mm). The distance between the probe tip and the tissue is stabilized at the imaging position with distance variance less than 40 μm. The area of mosaics for a spiral trajectory with the proposed control scheme is 8.61 mm². The proposed control scheme has a high potential to collect a clear pCLE image sequence by maintaining a constant distance between the probe tip and the tissue.
A study on using image serving technology for high precision mechanical positioning
2016, Mechanical Systems and Signal Processing
Citation Excerpt :
The features for image matching include color [18], texture [19–21], spatial relation [22], and shape [23]. In the past, studies have employed the gray-scale energy method [24], fast corner search method [25], and Fast Hough Transform [9]. Previously used matching cost computation [26] methods include: normalized cross-correlation (NCC), sum of squared differences (SSD), sum of absolute differences (SAD), and rank and census.
This paper focused on using image server technology for high precision mechanical positioning. Rapid and precise positioning systems depend on the correct positions of CCD (Charge Coupled Device) video cameras, as well as on pattern matching modes. This study deals with four different positions captured by an automatic detection system employing a CCD video camera. According to a variety of hybrid image registration systems, this study proposes an entire set of methods for achieving optimal hybrid pattern matching. First, the four different position detections captured by the CCD video camera in low-resolution were examined. Next, the original position detection was carried out in high-resolution, in order to derive a precise set of CCD video camera positions. The fiducial mark (FM) was then divided into two types in the optimal option for pattern matching: the “fiducial mark” and “non-fiducial mark”, which were then used for sampling. The automatic detection method is able to achieve the first pattern matching detection for recognized images. Unrecognized images or images that cannot have an FM were subjected to fine pattern matching detection. When it is not possible to find more than one FM after the proposed detection method, this suggests that the position of the CCD video camera should be reset. In this paper, the results of the experiment regarding the CCD camera precision location and the segmentation of fiducial patterns or insignificant fiducial patterns can detect and segment more unique areas and areas with unique features.
A new generalised α scale spaces quadrature filters
2014, Pattern Recognition
Citation Excerpt :
Also, it has been shown that this model successfully explains a number of psychophysical effects on human feature perception. This observation has led to the development of a number of phase-based feature detection algorithms [14,17,19,36–43]. Numerous additional advantages of the use of phase-based measurements have been reported.
The α scale spaces is a recent theory that open new possibilities of phase-based image processing. It is a parameterised class $(α \in] 0, 1])$ of linear scale space representations, which allows a continuous connection beyond the well-known Gaussian scale space ( $α = 1$ ). In this paper, we make use of this unified representation to derive new families of band pass quadrature filters, built from derivatives and difference of the α scale space generating kernels. This construction leads to generalised α kernel filters including the commonly known families derived from the Gaussian and the Poisson kernels. The properties of each family are first studied and then experiments on one- and two-dimensional signals are shown to exemplify how the suggested filters can be used for edge detection. This work is complemented by an experimental evaluation, which demonstrates that the new proposed filters are a good alternative to the commonly used Log-Gabor filter.
Edge and line oriented contour detection: State of the art
2011, Image and Vision Computing
Citation Excerpt :
In [111], the concept of constrained phase congruency is introduced, showing its ability to detect edges and other salient features such as Mach bands. In [112,113], local energy models are used for detecting a larger set of edge features, including junctions and line terminations. As discussed in [114], the ability of local energy models of detecting roof, line and Mach band edges makes this operator idempotent.
We present an overview of various edge and line oriented approaches to contour detection that have been proposed in the last two decades. By edge and line oriented we mean methods that do not rely on segmentation. Distinction is made between edges and contours. Contour detectors are divided in local and global operators. The former are mainly based on differential analysis, statistical approaches, phase congruency, rank order filters, and combinations thereof. The latter include computation of contour saliency, perceptual grouping, relaxation labeling and active contours. Important aspects are covered, such as preprocessing aimed to suppress texture and noise, multiresolution techniques, connections between computational models and properties of the human visual system, and use of shape priors. An overview of procedures and metrics for quantitative performance evaluation is also presented. Our main conclusion is that contour detection has reached high degree of sophistication, taking into account multimodal contour definition (by luminance, color or texture changes), mechanisms for reducing the contour masking influence of noise and texture, perceptual grouping, multiscale aspects and high-level vision information.
Comparative Analysis of Convolutional Neural Networks for Classification of Breast Abnormalities
2023, Journal of Communications Technology and Electronics

View all citing articles on Scopus

^∗: Present address: Systems Intellect, 1121 Hay St., West Perth 6005, Australia.

View full text

Regular paper2D feature detection via local energy

Abstract

Computer Vision, Graphics and Image Processing

Pattern Recognition Letters

Pattern Recognition Letters

Determination of ego-motion from matched points

Stereo correspondence: A hierarchical approach

Visual Perception of Form

Finding edges and lines in images

The curvature primal sketch

IEEE Trans. PAMI

Rotationally invariant image operators

A framework for low level feature extraction

Accurate corner detection: an analytical study

A computational approach to corner and vertex detection

International Journal of Computer Vision

Modelling and identification of characteristic intensity variations

Image and Vision Computing

Regular paper
2D feature detection via local energy