By Topic

Image Processing, IET

Issue 9 • Date December 2013

Filter Results

Displaying Results 1 - 6 of 6
  • Saliency-based localising active contour for automatic natural object segmentation

    Page(s): 787 - 794
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (869 KB)  

    In this study, a novel method named saliency-seeded active contour is presented for automatic natural object extraction. Since approximately the location of the desired object can easily be obtained by saliency regions or pixels in the map, we propose the maximum saliency density method to detect salient object pixels in spite of the cluttered background at first. Then, the salient object pixels are employed as the seeds of convex hull to generate the initial contour for our automatic object segmentation system. It is most important that the method proposed by the authors does not require considerable user interaction in contrast with localising region-based active contours (LRACs), that is, the segmentation task is fulfiled in a fully automatic manner. Extensive experiments results on a large variety of natural images confirm that the framework can reliably and automatically extract the object from the complex background. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive vectorial total variation models for multi-channel synthetic aperture radar images despeckling with fast algorithms

    Page(s): 795 - 804
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (818 KB)  

    This study proposes two adaptive vectorial total variation models for multi-channel synthetic aperture radar (SAR) images despeckling with the help of prior knowledge of the image amplitude. Besides despeckling the multi-channel SAR images efficiently, the proposed new models have advantages over other total variation methods in many aspects, such as preserving the radar reflectivity, the targets and edges contrast. The Bermudez-Moreno algorithm and the accelerated fast iterative shrinkage thresholding algorithm are employed to implement the new two models, respectively. Experimental results on multi-polarimetric, multi-temporal RADARSAT-2 images show that the visual quality and evaluation indexes of the proposed models and the corresponding algorithms outperform the other methods with edge preservation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Reversible data hiding based on two-dimensional prediction errors

    Page(s): 805 - 816
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (1137 KB)  

    The conventional histogram-based reversible data-hiding scheme is one-dimensional (1D). In this article, a novel framework that can be used to design 2D reversible data-hiding schemes is presented. Through the flexibility of selecting peaks for each channel, this 2D technique offers higher embedding performance than the conventional 1D techniques. For illustration, the framework is applied to develop two new schemes, C-2D and S-2D. Compared with the 1D schemes that use the predictions individually, the experimental results show that C-2D and S-2D have apparent performance advantages. This framework can be applied to any architecture. Furthermore, it can be easily extended into a multi-dimensional framework. By combining appropriate prediction methods, more reversible data-hiding schemes can be derived. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Forensic detection of image manipulation using the Zernike moments and pixel-pair histogram

    Page(s): 817 - 828
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (736 KB)  

    Integrity verification or forgery detection of an image is a difficult procedure, since the forgeries use various transformations to create an altered image. Pixel mapping transforms, such as contrast enhancement, histogram equalisation, gamma correction and so on, are the most popular methods to improve the objective property of an altered image. In addition, fabricators add Gaussian noise to the altered image in order to remove the statistical traces produced because of pixel mapping transforms. A new method is introduced to detect and classify four various categories including original, contrast modified, histogram-equalised and noisy images. In the proposed method, the absolute value of the first 36 Zernike moments of the pixel-pair histogram and its binary form for each image in the polar coordinates are calculated, and then those features that yield the maximum between-class separation, are selected. Some other features obtained from Fourier transform are also utilised for more separation. Finally, support vector machine classifier is used to classify the input image into four categories. The experimental results show that the proposed method achieves high classification rate and considerably outperforms the previously presented methods. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Pulse-coupled neural network feature generation model for Arabic sign language recognition

    Page(s): 829 - 836
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (450 KB)  

    Many feature generation methods have been developed for object recognition. Some of these methods succeeded in achieving invariance against object translation, rotation and scaling but faced problems of the bright background effect and non-uniform light on the quality of the generated features. This problem has hindered recognition systems from working in a free environment. This paper proposes a new method to enhance the feature quality based on pulse-coupled neural network. An adaptive model that defines continuity factor is proposed as a weight factor of the current pulse in signature generation process. The proposed new method has been employed in a hybrid feature extraction model that is followed by a classifier and was applied and tested in Arabic sign language static hand posture recognition; the superiority of the new method is shown. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Novel multifocus image fusion and reconstruction framework based on compressed sensing

    Page(s): 837 - 847
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (571 KB)  

    In this study, an efficient multifocus image fusion and reconstruction framework based on compressed sensing in the wavelet domain are proposed. The new framework is composed of three phases. Firstly, the source images are represented with their sparse coefficients using the discrete wavelet transform (DWT). Secondly, the measurements are obtained by the random Gaussian matrix from their sparse coefficients, and are then fused by the proposed adaptive local energy metrics (ALEM) fusion scheme. Finally, a fast continuous linearised augmented Lagrangian method (FCLALM) is proposed to reconstruct the sparse coefficients from the fused measurement, which will be converted by the inverse DWT (IDWT) to the fused image. Our experimental results show that the proposed ALEM image fusion scheme can achieve a higher fusion quality than some existing fusion schemes. In addition, the proposed FCLALM reconstruction algorithm has a higher peak-signal-to-noise ratio and a faster convergence rate as compared with some existing reconstruction methods. View full abstract»

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

Aims & Scope

The range of topics covered by IET Image Processing includes areas related to the generation, processing and communication of visual information.

Full Aims & Scope

Meet Our Editors

Publisher
IET Research Journals
iet_ipr@theiet.org