The Third International Conference on Signal and Image Processing (SIP 2014)

The contrast enhancement of medical images has an important role in diseases diagnostic, specially,cancer cases. Histogram equalization is considered as the most popular algorithm for contrast enhancement according to its effectiveness and simplicity. In this paper,we present a modified version of the Histogram Based Fast Enhancement Algorithm. This modified algorithm enhances the areas of interest with less complexity. It is applied only to CT head images and its idea based on treating with the soft tissues and ignoring other details in the image. The proposed modification make the algorithm is valid for most CT image types with enhanced results.

Bayesian networks are used today in more areas of decision support, classification and image analysis. The objective of this paper is to propose a new unsupervised Bayesian image segmentation. The proposed approach is to find the segmented image with the best overall quality, this quality is modeled by a Bayesian Network representing different causal relationships between the various elements of the image. This quality is calculated on a set of attributes that represent local evaluation measures which represent local levels of this segmentation, the idea is to have these local levels chosen in a way to be intersected into them to keep the overall appearance of segmentation. The approach operates in two phases: the first phase is to make an over-segmentation which gives superpixels card. In the second phase, we model the superpixels by a Bayesian Network. An approximate inference algorithm then allows to find the segmented image with the best overall quality. For this purpose, we used two approximate inference methods, the first using ICM algorithm which is widely used in Markov Models and a second is a recursive method called algorithm of model decomposition based on max-product algorithm which is very popular in the recent works of image segmentation. For our model, we have shown that the composition of these two algorithms leads to good segmentation performance.

In this paper we implemented an Auto-exposure algorithm based on the Falseposition method in order to correctly expose the leather samples. Though we are doing this for the leather industry, we can directly use this false position based, auto exposure algorithm for the natural scenes. The main reason for choosing the Falseposition method is that, it converge the root values quickly when compared to bisection method. The implementation of our auto exposure algorithm is performed by using the point grey research programmable camera.

Real time road-marking recognition has a wide variety of application from assisting driver, autonomous driving to helping partially sighted to walk on road. This area was touched very slightly till now. Our paper presents a robust real-time approach which can detect all types of road markings, pictograms,symbols, texts etc. Our algorithm is based on the following steps: converting image from RGB to HSL color space, applying thresholding to detect required pixels, remove noise and outliers using Connected Component Analysis followed by Random Sample Consensus Approach. The approach works well in different weather conditions, road conditions and quality of images.

The fingerprint images compression based on geometric transformed presents important research topic, these last year's many transforms have been proposed to give the best representation to a particular type of image "fingerprint image", like classics wavelets and wave atoms. In this paper we shall present a comparative study between this transforms, in order to use them in compression. The results show that for fingerprint images, the wave atom offers better performance than the current transform based compression standard. The wave atoms transformation brings a considerable contribution on the compression of fingerprints images by achieving high values of ratios compression and PSNR, with a reduced number of coefficients. In addition, the proposed method is verified with objective and subjective testing

Document images often suffer from different types of degradation that renders the document image binarization a challenging task. In this paper, a new binarization algorithm for degraded document images is presented. The method is based on active contours evolving according to intrinsic geometric measures of the document image; Niblack's thresholding is also used to control the active contours propagation. The validity of the proposed method is demonstrated on both recent and historical document images including different types of degradations, the results are compared with a number of known techniques in the literature.

Uncertainty is an inherent part of intelligent systems used in face recognition applications. The use of new methods for handling inaccurate information regarding facial features is of fundamental importance. This paper deals with the design of intelligent 2D face recognition system using interval type-2 fuzzy logic for minimizing the effects of uncertainty produced by variations in light direction, face pose and facial expression. Built on top of the well-known fisherface method, our system employs type-2 fuzzy set to compute fuzzy within and in-between class scatter matrices of fisher's linear discriminant. This employment makes the system able to improve face recognition rates as the results of reducing the sensitivity to substantial variations between face images. Type-2 Fuzzy Sets (T2FSs) have been shown to manage uncertainty more effectively than Type-1 Fuzzy Sets (T1FS), because they provide us with more parameters that can handle environments where it is difficult to define an exact membership function for a fuzzy set. Experimental results for Yale face databases are given, which show the effectiveness of the suggested system for face recognition and also show the privilege and high accuracy when compared with other methods.

It was observed that for non-stationary and quasi-stationary signals, wavelet transform has been found to be an effective tool for the time-frequency analysis. Recent years have seen wavelet transform being used for feature extration in speech recognition applications. Here a new filter structure using admissible wavelet packet analysis is proposed. These filters have the benefit of having frequency bands spacing similar to the auditory gammatone filter whose central frequencies are equally distributed along the ERB scale. A new sets of features have been derived using wavelet packet transform's multi-resolution capabilities, which perform better than conventional features like GFCC and MFCC for unvoiced phoneme recognition problems.

The actual impacts of different enhancement and display methods on the data observations are difficult to measure, but they can provide invaluable information to the software and algorithm developers. In this study we analyse the usability of saliency detection algorithms as a tool to capture the impacts of different data enhancement and display methods. We capture the interpreters' eye gaze movements during target spotting exercises within magnetic datasets, a type of geophysical image commonly used in resource exploration. The ITTI saliency detection algorithm highlighted the areas similar to the interpreter data observation patterns and from this we concluded this best represents the human visual attention. This algorithm was then used to assess a suite of images to identify the most effective image enhancement and display methods. Our results show several common forms of display produced similar saliency maps and as such may not all be needed to complete an interpretation.

One of the goals of Steganography is to design schemes with high embedding efficiency, which can be broadly defined as the ratio between the amount of the communicated information and the amount of introduced distortion. In this contribution, a novel method of steganographic embedding is described by vertex coloring in the graphs product. We investigate the cartesian product for improvement CI-rate of ±1 Steganography. Finally, the performance of the proposed methods is evaluated and compared with those of the aforementioned methods.

The ability of communicating easily with everyone is a blessing people with hearing impairment do not have. They completely rely on their vision around healthy individuals to difficultly read lips. This paper proposes a solution for this problem, ISWYS (I See What You Say) is a research-oriented speech recognition system for the Arabic language that interprets lips movements into readable text. It is accomplished by analyzing a video of lips movements that resemble utterances, and then converting it to readable characters using video analysis and motion estimation. Our algorithm involves dividing the video into n number of frames to generate n-1 image frame which is produced by taking the difference between consecutive frames. Then video features are extracted to be used by our error function which provided a recognition of approximately 70%.

In this paper, we proposed a segmentation approach that not only segment an interest object but also label different semantic parts of the object, where a discriminative model is presented to describe an object in real world images as multiply, disparate and correlative parts. We propose a multi-stage segmentation approach to make inference on the segments of an object. Then we train it under the latent structural SVM learning framework. Then, we showed that our method boost an average increase of about 5% on ETHZ Shape Classes Dataset and 4% on INRIA horses dataset. Finally, extensive experiments of intricate occlusion on INRIA horses dataset and robot experiments show that the approach have a state of the art performance in the condition of occlusion and deformation.

Realistic simulation of natural phenomena such as clouds is one of the most challenging problems facing computer graphics. The complexity of cloud formation, dynamics and light interaction makes real time cloud rendering a difficult task. In addition, traditional computer graphics methods involve high amounts of memory and computing resources, which currently limits their realism and speed. We propose an efficient and computationally inexpensive phenomenological approach for modelling and rendering cumulus clouds, by drawing on several approaches that we combine and extend.This paper focuses on the modelling of the cloud's shape, rendering and sky model but is does not deal with the animation of the cloud.

Gait recognition is kind of biometric technology that can be used to monitor people without their cooperation. Controlled environments such as banks, military installations and even airports need to be able to quickly detect threats and provide differing levels of access to different user groups. Gait shows a particular way or manner of moving on foot and gait recognition is the process of identifying an individual by the manner in which they walk. Gait is less unobtrusive biometric, which offers the possibility to identify people at a distance, without any interaction or co-operation from the subject; this is the property which makes it so attractive.

In this paper, first step is extraction of foreground objects i.e. human and other moving objects from input video sequences or binary silhouette of a walking person is detected from each frame and human detection and tracking will be performed. Next, feature extraction process is used to extract the gait features of a person. Vector of outer contour of binary silhouette and MPEG-7 ART (Angular Radial Transform) coefficients are taken as the feature vector. At last BPNN(back propagation neural network) technique is used for training and testing purpose and percentage of matching will be calculated which is far better than previous research. Here all experiments are done on gait database and input video.

In this paper, we use the bilateral filter to de-noise nighttime infrared monitoring images, in order that noises are reduced and edges are preserved. Considering that traditional bilateral filter does not balance gray-scale difference and the geometric distance in determining the weights of convolution kernel coefficients, we propose to give the two different values different weights, so that both of them are fully considered. Experimental results show that this algorithm can effectively remove noise as well as perfectly retain edge information.

Most watermark methods use pixel values or coefficients as the judgment condition to embed or extract a watermark image. The variation of these values may lead to the inaccurate condition such that an incorrect judgment has been laid out. To avoid this problem, we design a stable judgment mechanism, in which the outcome will not be seriously influenced by the variation. The principle of judgment depends on the scale relationship of two pixels. From the observation of common signal processing operations, we can find that the pixel value of processed image usually keeps stable unless an image has been manipulated by cropping attack or halftone transformation. This can greatly help reduce the modification strength from image processing operations. Experiment results show that the proposed method can resist various attacks and keep the image quality friendly.

The Phonocardiogram (PCG) is the graphical representation of acoustic energy due to the mechanical cardiac activity. Sometimes cardiac diseases provide pathological murmurs mixed with the main components of the Heart Sound Signal (HSs). The pathological murmurs may be associated with heart disease or valvular heart disease. The Empirical Mode Decomposition (EMD) allows decomposing a multicomponent signal into a set of monocomponent signals, called Intrinsic Mode Functions (IMFs). Each IMF represents a monocomponent signal or an oscillatory mode with one instantaneous frequency. The goal of this paper is to segment some pathological HSs by selecting the most appropriate IMFs characterizing the murmur using the correlation coefficient. The experimental results conducted on some real-life pathological heart sound signals such as: mitral regurgitation, aortic regurgitation and the opening snap case; revealed the performance of the proposed method. This method is simple and gives best results in comparison with the known method based on the noise only model which requires the use of a model for his implementation.

This paper proposes a neural network based region classification technique that classifies regions in an image into two classes: textures and homogenous regions. The classification is based on training a neural network with statistical parameters belonging to the regions of interest. An application of this classification method is applied in image denoising by applying different transforms to the two different classes. Texture is denoised by shearlets while homogenous regions are denoised by wavelets. The denoised results show better performance than either of the transforms applied independently. The proposed algorithm successfully reduces the mean square error of the denoised result and provides perceptually good results.

In this paper we present the Large Inverse Cholesky (LIC) method for efficient computing of the coefficient matrices of a Structured Vector Autoregressive model.

Emotional state recognition through speech is being a very interesting research topic nowadays. Using subliminal information of speech, it is possible to recognize the emotional state of the person. One of the main problems in the design of automatic emotion recognition systems is the small number of available patterns. This fact makes the learning process more difficult, due to the generalization problems that arise under these conditions.
In this work we propose a solution to this problem consisting in enlarging the training set through the creation the new virtual patterns. In the case of emotional speech, most of the emotional information is included in speed and pitch variations. So, a change in the average pitch that does not modify neither the speed nor the pitch variations does not affect the expressed emotion. Thus, we use this prior information in order to create new patterns applying a pitch shift modification in the feature extraction process of the classification system. For this purpose, we propose a frequency scaling modification of the Mel Frequency Cepstral Coefficients, used to classify the emotion. This proposed process allows us to synthetically increase the number of available patterns in thetraining set, thus increasing the generalization capability of the system and reducing the test error.

The paper describes the image acquisition system able to capture images in two separated bands of light,used to underwater autonomous navigation. The channels are: the visible light spectrum and near infrared spectrum. The characteristics of natural, underwater environment were also described together with the process of the underwater image creation. The results of an experiment with comparison of selected images acquired in these channels are discussed.

Time Of Flight Diffraction (TOFD) is a non destructive testing technique that analyses the diffraction of the ultrasonic waves by the tips of discontinuities.In this technique, data are displayed in the form of images and processing algorithms can be applied, allowing automatic detection and characterisation of the discontinuities presented in the material. When large surfaces are inspected, the amount of data to process is considerable. This makes suitable the limitation of records and the development of algorithms to give help in decision making. The work presented here is a data processing algorithm which allows to detect and to locate automatically cracks in a structure during a TOFD technique inspection. An original approach to data storage is proposed in order to avoid image formation, replacing it by a sparse matrix, The application of the RAndom SAmple and Consensus (RANSAC) filter is then exploited to automate the recognition of crack defects by detecting parabolic forms formed by the matrix elements in this case.