Idris El-Feghi

X-ray image segmentation using auto adaptive fuzzy index measure

Image segmentation is a crucial step in a wide range of medical image processing systems. It is useful in visualization of the different objects present in the image. For example separation of the soft, boney tissues and background on the lateral skull X-ray plays an important role in producing cephalometric tracing and hence producing accurate cephalometric evaluation used in orthodontic practice. In spite of the several methods available in the literature, image segmentation still a challenging problem in most of the image processing applications. The challenge comes from the fuzziness of image objects and the overlapping of the different regions. In this paper we propose fast auto adaptive image segmentation algorithm for finding the optimal thresholds for segmenting gray scale images. The proposed method is based on minimizing a fuzzy index which decreases as the similarity between pixels increases. The system uses initial estimates of the parameters of the fuzzy subsets derived from the image histogram then uses fuzzy entropy as cost measure to maximize the similarity between pixels of the same subset. Experimental results demonstrate the effectiveness of the proposed approach.

An adaptive ant-based clustering algorithm with improved environment perception

Data clustering plays an important role in many disciplines, including data mining, machine learning, bioinformatics, pattern recognition, and other fields. When there is a need to learn the inherent grouping structure of data in an unsupervised manner, ant-based clustering stand out as the most widely used group of swarm-based clustering algorithms. Under this perspective, this paper presents a new Adaptive Ant-based Clustering Algorithm (AACA) for clustering data sets. The algorithm takes into account the properties of aggregation pheromone and perception of the environment together with other modifications to the standard parameters that improves its convergence. The performance of AACA is studied and compared to other methods using various patterns and data sets. It is also compared to standard clustering using a set of analytical evaluation functions and a range of synthetic and real data collection. Experimental results have shown that the proposed modifications improve the performance of ant-colony clustering algorithm in term of quality and run time.

Automatic localization of craniofacial landmarks for assisted cephalometry

In this work we propose a system for localization of cephalometric landmarks. The process of localization is carried out in two step: deriving a smaller expectation window for each landmark using a trained neuro-fuzzy system (NFS,) then applying a template-matching algorithm to pin point the exact location of the landmark. The system is trained to locate 20 landmarks on a database of 200 images. Preliminary results show an average of 90% recognition rate.

Craniofacial landmarks extraction by Partial Least Squares Regression

In this paper, a novel method based on Partial Least Square Regression (PLSR) is introduced to extract the relation between selected point coordinates on X-ray images and the expected location of a set of landmarks formally known as craniofacial landmarks. In the proposed method, four points are located using image detection techniques. The four points are used to extract more features representing rotation, scale and shift in the lateral skull X-ray. PLSR is used to obtain a predictive matrix for the landmarks on the tests set. Experimental results showed that this method could locate landmarks with an accuracy of 75%.

Content-Based Image Retrieval based on efficient fuzzy color signature

Content-Based Image Retrieval (CBIR) Systems are becoming increasingly important and finding application in diverse areas of discipline. Due to disturbing factors present in images, e.g. non-uniform illumination and lightening intensities, the retrieval results are not always satisfactory. In this paper, we discuss how fuzzy set theory can be used to formulate an efficient image signature. This signature is 1-D Fuzzy Color Histogram (FCH) consisting of a small number of bins. The signature is obtained from image contents by considering the contribution of each pixels color to all the histogram bins through the use of fuzzy-sets membership functions. Experimental results on a database of over 10,000 images demonstrate that the proposed system is less sensitive light intensity changes and more robust than the Conventional Color Histogram (CCH) in retrieval precision.

Improved Co-occurrence Matrix as a Feature Space for Relative Entropy-based Image Thresholding

In this paper, a thresholding technique suitable for noisy background images is proposed. The proposed algorithm uses an improved co-occurrence matrix as feature spaces. The threshold value is obtained by maximizing the relative entropy. Experimental results show that the proposed method outperforms other thresholding techniques especially on the presences of noise in the background of the input image.

A binarization method for scanned documents based on hidden Markov model

The binarization stage is a very critical in document analysis since the quality of a binarized image determines the performance of the entire process. However due to many uncertain factors such as complex signal-dependent noise and variable background intensity, which is caused by non-uniform illumination, shadow, smear, smudge or low contrast, an effective binarization algorithm that will work on a broad set of composite documents has yet to be developed. Here we propose a new algorithm which implements hidden Markov model in the binarization stage. The simulation results prove that this method is efficient in extracting information from various noisy backgrounds

Automatic localization of craniofacial landmarks using multi-layer perceptron as a function approximator

There are 20-30 visible landmarks in the lateral X-ray skull that are used by orthodontists in what is known as cephalometric evaluation. The evaluation assists in the diagnosis of anomalies and in the monitoring of treatments. Currently, this process is carried-out manually by outlining the soft and bonny tissues of the skull then locating the landmarks on line crossings. This can take an experienced orthodontist up to 20min. The process is tedious, time consuming and subject to human error. In this paper, we propose a system for automatic localization of cephalometric landmarks using Multi-Layer Perceptron (MLP). Image processing techniques are utilized to extract features representing rotation, scale and distances from the outer edges of the skull image. Features from manually labeled images are used as inputs to train the MLP. After training, the MLP is used to estimate the location of the landmarks on targeted images based on knowledge obtained on the training stage. Results obtained by testing the algorithm on images which are not seen by the MLP during training, show an improvement over previously reported techniques.

Contrast enhancement of radiograph images based on local heterogeneity measures

Contrast enhancement of the lateral skull X-ray images is very important in orthodontic practice, cephalometric evaluation and craniofacial landmarking. In this paper we present computational techniques involving contrast enhancement of two-dimensional digitized X-ray images. The algorithm is developed based on the use of local heterogeneity measures of pixel distribution. Image enhancement is accomplished by an adaptive gray scale pixels transformation depending on results of local heterogeneity contribution measures. The proposed approach was tested on different images and the results prove that the proposed method has better performance the existing conventional methods.

Automatic Identification and Localization of Craniofacial Landmarks Using Multi Layer Neural Network

Cephalometric evaluation of lateral x-rays of the skull, used mainly by orthodontists, is usually carried-out manually to locate certain craniofacial landmarks. This process is time consuming, which is both tedious and subject to human error. In this paper we propose a novel algorithm based on the use of the Multi-layer Perceptron (MLP) to locate landmarks on the digitized x-ray of the skull. The main feature of this proposed algorithm is that its performance is independent of the quality of radiographs. Preprocessing techniques are used to enhance the quality of the image and to extract the outer edges of the skull. Four points are selected to form the basis for additional features representing the size, rotation and offset of the skull. The extracted features are then used as inputs to the MLP. The corresponding outputs represent the horizontal and vertical coordinates of the selected landmark. MLP’s are efficient function approximators and in this work are trained to locate landmarks by using a number of manually labeled data as a training set. After training, the MLP is used to locate landmarks on target digitized images of radiographs. The MLP is trained using 55 manually labeled images and tested on a separate set consisting of 134 images, which are not used for training. Results obtained show an improvement over template-matching and line-following techniques. This is apparently evident when the search encounters a lost tooth, cavity filling or when the image is of a low quality.