Jan Nordin

Indoor navigation to support the blind person using weighted topological map

This paper introduces a new approach of an electronic cane for navigation through the environment, using only vision system to help the blind people. By computing weights between already stored images and the real scene of the environment. The system gives advices for the blind person to select the right direction. This advice depends on a weighted topological map in the form of an appearance graph. Navigation on this graph involves Importance value from one node to other until the goal node is reached. Where a mono camera cane-held gives information in front of the blind person. The system will give a wide range indoor navigation and may be used to outdoor navigation. The identification of different scenes to the blind person has done with in a session. These sessions divide the image database into parts like bedroom, corridor,….etc. The proposed scheme employs SIFT features to represent scene containing many objects in the environment.

A Chaotic Cryptosystem for Images Based on Henon and Arnold Cat Map

The rapid evolution of imaging and communication technologies has transformed images into a widespread data type. Different types of data, such as personal medical information, official correspondence, or governmental and military documents, are saved and transmitted in the form of images over public networks. Hence, a fast and secure cryptosystem is needed for high-resolution images. In this paper, a novel encryption scheme is presented for securing images based on Arnold cat and Henon chaotic maps. The scheme uses Arnold cat map for bit- and pixel-level permutations on plain and secret images, while Henon map creates secret images and specific parameters for the permutations. Both the encryption and decryption processes are explained, formulated, and graphically presented. The results of security analysis of five different images demonstrate the strength of the proposed cryptosystem against statistical, brute force and differential attacks. The evaluated running time for both encryption and decryption processes guarantee that the cryptosystem can work effectively in real-time applications.

Review on statistical approaches for automatic image annotation

This paper will discuss the literature survey on statistical approaches for automatic annotation on digital images. There are some researches on image annotation and produced very good knowledge theoretically or technically and lead to produce such promising surveys. However, not all produced related knowledge on image annotation and discussing until the latest research on that area. This paper aims to cover the statistical approaches for automatic image annotation from it starts until the latest research finding. Using the earlier literature reviews, approaches, frameworks and evaluation results as guidelines, an attempt is made to outline the automatic image annotation which is combination of image analysis and statistical learning approaches. Summary and analysis of some of the approaches have been used as references to produce a framework in designing an automatic image annotation model.

Combining Local Binary Pattern and Principal Component Analysis on T-Zone face area for face recognition

This paper presents a combination techniques of appearance-based and feature-based feature extraction on the T-Zone face area to improve the recognition performance. This study shows that the T-Zone area and the combined technique provides a significant impact on the face recognition rate. A T-Zone face image is first divided into small regions where Local Binary Pattern (LBP) histograms are extracted and then concatenated into a single feature vector. This feature vector will further reduce the dimensionality scope by using the well established Principle Component Analysis (PCA) technique. Experiments have been carried out on the different sets of the Olivetti Research Laboratory (ORL) database. High recognition rates are obtained when compared to other face recognition methods of the same class. Our result shows of 7% improvement compared with PCA and 2% improvement compare with Sub-Holistic PCA. Our studies proves that the T-Zone area which is consisting of eyes and nose region is a ‘significant facial region’, and we also show that LBP can easily be combined with PCA to reduce the length of the feature vector, while the recognition performance is improved.

SIFT based monocular SLAM with multi-clouds features for indoor navigation

This work introduces a monocular SLAM method, which uses the Scale Invariant Features Transform (SIFT) representation for the scene. The scene represented as clouds of SIFT features within the map. This hierarchical representation of space, serving to estimate the current direction in the environment within the current session. The system exploits the tracking of the same features of successive frames to calculate scalar weights for these features, to build a map of the environment indicating the camera movement, helping the blind persons to navigate more confidently through auditory pathway of their surroundings. EKF is used to estimate the features tracked within the successive frames. The system is tested for using the proposed method with a hand-held camera walking in indoor environment. The results show a good estimation on the spatial locations of the camera within a few milliseconds. The paper shows an electronic cane for navigating in indoor environment using these clouds of features for long-term appearance-based localization of a cane with web camera vision as the external sensor.

Invariant Feature Matching for Image Registration Application Based on New Dissimilarity of Spatial Features

An invariant feature matching method is proposed as a spatially invariant feature matching approach. Deformation effects, such as affine and homography, change the local information within the image and can result in ambiguous local information pertaining to image points. New method based on dissimilarity values, which measures the dissimilarity of the features through the path based on Eigenvector properties, is proposed. Evidence shows that existing matching techniques using similarity metrics—such as normalized cross-correlation, squared sum of intensity differences and correlation coefficient—are insufficient for achieving adequate results under different image deformations. Thus, new descriptor’s similarity metrics based on normalized Eigenvector correlation and signal directional differences, which are robust under local variation of the image information, are proposed to establish an efficient feature matching technique. The method proposed in this study measures the dissimilarity in the signal frequency along the path between two features. Moreover, these dissimilarity values are accumulated in a 2D dissimilarity space, allowing accurate corresponding features to be extracted based on the cumulative space using a voting strategy. This method can be used in image registration applications, as it overcomes the limitations of the existing approaches. The output results demonstrate that the proposed technique outperforms the other methods when evaluated using a standard dataset, in terms of precision-recall and corner correspondence.

Regression techniques for the prediction of stock price trend

This paper examines the theory and practice of regression techniques for prediction of stock price trend by using a transformed data set in ordinal data format. The original pretransformed data source contains data of heterogeneous data types used for handling of currency values and financial ratios. The data formats in currency values and financial ratios provide a process for computation of stock prices. The transformed data set contains only a standardized ordinal data type which provides a process to measure rankings of stock price trends. The outcomes of both processes are examined and appraised. The primary design is based on regression analysis from WEKA machine learning software. The stock price movement in Bursa Malaysia is used as our research setting. The data sources are corporate annual reports which included balance sheet, income statement and cash flow statement. The variables included in the data set were formed based on stock market trading fundamental analysis approach. Classifiers in WEKA were used as algorithms to produce the outcomes. This study showed that the outcomes of regression techniques can be improved for the prediction of stock price trend by using a dataset in standardized ordinal data format.

Deformation invariant image matching based on dissimilarity of spatial features

In this paper, a new deformation invariant image matching method, known as spatial orientation feature matching (SOFM), is presented. A new similarity value, which measures the similarity of the signal through the path based on triple-wise signal eigenvector correlation, is proposed. The proposed method extracts similarity feature values by relying on the distinct path between two specific interest points and following the alternation of the signal while traversing the path. Because these similarity values of the path are deformation invariant, the proposed method supports various types of transformation in the original image, such as scale, translation, rotation, intensity noises and occlusion. Moreover, the triple-wise similarity scores are accumulated in a 2-D similarity space; thus, robust matched correspondence points are obtained using cumulative similarity space. SOFM was compared to the most recent related methods using corner correspondence (CC) and precision-recall evaluation metrics. The findings confirmed that SOFM provides higher correspondence ratios, and the results indicate that it outperforms currently utilized methods in terms of accuracy and generalization.

Bee royalty offspring algorithm for improvement of facial expressions classification model

A major issue which divides the facial expressions from the other classification domains is complicated behaviour of human to express the emotions which should be recognised with the classifier model. Existing research recognise the emotions using a range of classification techniques. However, low accuracy rate, large training set, large extracted features or priority for sequence images are the main drawbacks of those works. One of the recent techniques to address the facial expressions problem is fuzzy rule-based system FRBS which is used as a successful method to model and solve the natural-based problems. However, FRBS is poor to adapt the existing knowledge with the diverse conditions. In this article a novel hybrid genetic-fuzzy rule-based model is proposed to optimise the performance of fuzzy classification while the limited raw input data as the features are used. In this model, the proposed genetic algorithm simulates the honey bees offspring generation process called bee royalty offspring algorithm BROA to improve the training process of classic genetic algorithm. The comparison results illustrated that the genetic-fuzzy classification model improves considerably the accuracy rate and performance of FRBS while the BROA modify the training process of genetic-based algorithms.

Vision-based automatic incident detection system using image sequences for intersections

Traffic incident detection is one of the interesting fields of intelligent transportation system (ITS) which recently rapidly increasing interest in their used. In this paper, we proposed an incident detection system based on incident features and reporting traffic incident in a special intersection using machine vision algorithms. The first step in this algorithm after image sequences acquisition from the video image of CCD camera is vehicle detection. Then the incident features such as direction of the moving vehicles, traffic flow and the rate of changing speed will extract in order to achieve the detection results. Machine vision based algorithm has been used in order to develop the system for incident detection goal. This process gives the best result by total 97.8% of correct rate, 1.02 of false alarm rate and 30(S) is the meantime to detect. The result shows that this algorithm has a good detection rate.