By Topic

Computer Vision, IET

Issue 2 • Date March 2011

Filter Results

Displaying Results 1 - 7 of 7
  • Improving Harris corner selection strategy

    Publication Year: 2011 , Page(s): 87 - 96
    Cited by:  Papers (2)
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (1151 KB)  

    This study describes a corner selection strategy based on the Harris approach. Corners are usually defined as interest points for which intensity variation in the principal directions is locally maximised, as response from a filter given by the linear combination of the determinant and the trace of the autocorrelation matrix. The Harris corner detector, in its original definition, is only rotationally invariant, but scale-invariant and affine-covariant extensions have been developed. As one of the main drawbacks, corner detector performances are influenced by two user-given parameters: the linear combination coefficient and the response filter threshold. The main idea of the authors' approach is to search only the corners near enhanced edges and, by a z-score normalisation, to avoid the introduction of the linear combination coefficient. Combining these strategies allows a fine and stable corner selection without tuning the method. The new detector has been compared with other state-of-the-art detectors on the standard Oxford data set, achieving good results showing the validity of the approach. Analogous results have been obtained using the local detector evaluation framework on non-planar scenes by Fraundorfer and Bischof. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Polygonal shape reconstruction in the plane

    Publication Year: 2011 , Page(s): 97 - 106
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (999 KB)  

    In this study a robust shape reconstruction algorithm is proposed which guarantees a simple polygon as output and works well on both types of input, dot patterns and boundary samples, in the plane. Guaranteed polygonal output makes it favourable for many applications because of its ease of manipulation and use. The proposed algorithm, called simple-shape, starts reconstruction from the convex hull and makes it concave step by step based on a new hybrid selection criterion which is built on human beings visual perception. Also at the end, a simple-shape algorithm is tested in several cases and the results are compared with the latest shape reconstruction algorithm in the literature. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Face recognition using regularised generalised discriminant locality preserving projections

    Publication Year: 2011 , Page(s): 107 - 116
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (906 KB)  

    Discriminant locality preserving projection (DLPP) is a recently proposed algorithm, which is an extension of locality preserving projections (LPP) and can encode both the geometrical and discriminant structure of the data manifold. However, DLPP suffers from small sample size (SSS) problem which is often encountered in face recognition tasks. To deal with this problem, the authors propose a novel regularised generalised discriminant locality preserving projections (RGDLPP) method for facial feature extraction and recognition. First, locality preserving within-class scatter in DLPP method is replaced by locality preserving total scatter and all the training samples are projected into the range of locality preserving total scatter. Then the authors regularise the small and zero eigenvalues of locality preserving within-class scatter since the small eigenvalues are sensitive to noise. RGDLPP address SSS problem by removing the null space of locality preserving total scatter without loss of discriminant information. Meanwhile, RGDLPP can alleviate the problem of noise disturbance of the small eigenvalues. Experiments on the ORL, Yale, FERET and PIE face databases show the effectiveness of the proposed RGDLPP. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Three-dimensional machine vision and machinelearning algorithms applied to quality control of percussion caps

    Publication Year: 2011 , Page(s): 117 - 124
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (419 KB)  

    The exhaustive quality control is becoming very important in the world́s globalised market. One example where quality control becomes critical is the percussion cap mass production, an element assembled in firearm ammunition. These elements must achieve a minimum tolerance deviation in their fabrication. This study outlines a machine vision system development using a three-dimensional camera for the inspection of the whole production of percussion caps. This system presents multiple problems, such as metallic reflections in the percussion caps, high-speed movement for scanning the pieces, and mechanical errors and irregularities in percussion cap placement. Owing to these problems, it is impossible to solve the problem using traditional image processing methods, and hence, machine-learning algorithms have been tested to provide a feasible classification of the possible errors present in the percussion caps. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Embedded implementation of image-based water-level measurement system

    Publication Year: 2011 , Page(s): 125 - 133
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (1096 KB)  

    This paper proposes an embedded system for measuring the water level on a river using a camera. One of the main problems of using a camera in water-level measurement is that the light reflection on water surface changes rapidly. To solve this problem, this paper uses the property that variation of light reflection on the ruler showing the scale is relatively smaller than on the water surface. To emphasise this difference of light reflectance on the ruler surface and water surface, the proposed method accumulates the degree of light reflectance variation between the current image and the previous images and constructs a histogram using them. In this histogram, the boundary between the ruler and water surface is appeared as the sharpest drop point. To take into account the noise effects, the proposed algorithm finds the candidate points first and then finally select the water level among them using the reference water. The reference water level is roughly determined as the sharpest drop point in the histogram of the binary image. By installing the proposed system on actual river and testing the water-level measuring experiment in real time, its efficiency has been confirmed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Structured learning approach to image descriptor combination

    Publication Year: 2011 , Page(s): 134 - 142
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (411 KB)  

    In this study, the authors address the problem of combining descriptors for purposes of object categorisation and classification. The authors cast the problem in a structured learning setting by viewing the classifier bank and the codewords used in the categorisation and classification tasks as random fields. In this manner, the authors can abstract the problem into a graphical model setting, in which the fusion operation is a transformation over the field of descriptors and classifiers. Thus, the problem reduces itself to that of recovering the optimal transformation using a cost function which is convex and can be converted into either a quadratic or linear programme. This cost function is related to the target function used in discrete Markov random field approaches. The authors demonstrate the utility of our algorithm for purposes of image classification and learning class categories on two datasets. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast semi-global stereo matching via extracting disparity candidates from region boundaries

    Publication Year: 2011 , Page(s): 143 - 150
    Cited by:  Papers (2)
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (666 KB)  

    This study proposes a novel fast stereo matching algorithm via semi-global energy optimisation, which achieves a considerable improvement in efficiency for just a small price in accuracy. Based on some assumptions, the authors discover that at most two disparity candidates for each scanline segment of reference image can be extracted. With this observation, the authors present a disparity candidate extraction algorithm. This algorithm constructs an energy function based on colour consistency and restrictions between region boundaries. In this approach, the energy function is optimised via the graph-cuts technique, and the pixels involved are only those positioned on region boundaries, which results in greatly reduced vertex number in the constructed graph and subsequently improved efficiency. After that, a simple partial occlusions handling is conducted as a post-processing to enhance the accuracy of the final disparity map, by selecting a right disparity for each segment from extracted candidates. The performances of our method are demonstrated by experiments on the Middlebury test set. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.

Aims & Scope

IET Computer Vision seeks original research papers in a wide range of areas of computer vision. The vision of the journal is to publish the highest quality research work that is relevant and topical to the field, but not forgetting those works that aim to introduce new horizons and set the agenda for future avenues of research in Computer Vision.

Full Aims & Scope