Suphakant Phimoltares

Face detection and facial feature localization without considering the appearance of image context

Face and facial feature detection plays an important role in various applications such as human computer interaction, video surveillance, face tracking, and face recognition. Efficient face and facial feature detection algorithms are required for applying to those tasks. This paper presents the algorithms for all types of face images in the presence of several image conditions. There are two main stages. In the first stage, the faces are detected from an original image by using Canny edge detection and our proposed average face templates. Second, a proposed neural visual model (NVM) is used to recognize all possibilities of facial feature positions. Input parameters are obtained from the positions of facial features and the face characteristics that are low sensitive to intensity change. Finally, to improve the results, image dilation is applied for removing some irrelevant regions. Additionally, the algorithms can be extended to rotational invariance problem by using Radon transformation to extract the main angle of the face. With more than 1000 images, the algorithms are successfully tested with various types of faces affected by intensity, occlusion, structural components, facial expression, illumination, noise, and orientation.

Food Recommendation System Using Clustering Analysis for Diabetic Patients

Food and nutrition are a key to have good health. They are important for everyone to maintain a healthy diet especially for diabetic patients who have several limitations. Nutrition therapy is a major solution to prevent, manage and control diabetes by managing the nutrition based on the belief that food provides vital medicine and maintains a good health. Typically, diabetic patients need to avoid additional sugar and fat so the food pyramid is recommended to the patients for finding the substitution from the same food group. However, there is still a dietary diversity within food groups that can affect the diabetic patients. In this study, we proposed Food Recommendation System (FRS) by using food clustering analysis for diabetic patients. Our system will recommend the proper substituted foods in the context of nutrition and food characteristic. We used Self-Organizing Map (SOM) and K-mean clustering for food clustering analysis which is based on the similarity of eight significant nutrients for diabetic patient. At the end, the FRS was evaluated by nutritionists and it has performed very well and useful for nutrition area.

A Very Fast Neural Learning for Classification Using Only New Incoming Datum

This paper proposes a very fast 1-pass-throw-away learning algorithm based on a hyperellipsoidal function that can be translated and rotated to cover the data set during learning process. The translation and rotation of hyperellipsoidal function depends upon the distribution of the data set. In addition, we present versatile elliptic basis function (VEBF) neural network with one hidden layer. The hidden layer is adaptively divided into subhidden layers according to the number of classes of the training data set. Each subhidden layer can be scaled by incrementing a new node to learn new samples during training process. The learning time is O(n), where n is the number of data. The network can independently learn any new incoming datum without involving the previously learned data. There is no need to store all the data in order to mix with the new incoming data during the learning process.

3D CAPTCHA: A Next Generation of the CAPTCHA

Nowadays, the Internet is now becoming a part of our everyday lives. Many services, including Email, search engine, and web board on Internet, are provided with free of charge and unintentionally turns them into vulnerability services. Many software robots or, in short term, bots are developed with purpose to use such services illegally and automatically. Thus, web sites employ human authentication mechanism called Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA) to counter this attack. Unfortunately, many CAPTCHA have been already broken by bots and some CAPTCHA are difficult to read by human. In this paper, a new CAPTCHA method called 3D CAPTCHA is proposed to provide an enhanced protection from bots. This method based on assumption that human can recognize 3D character image better than Optical Character Recognition (OCR) software bots.

Facial expression recognition using graph-based features and artificial neural networks

Facial expression is significant for face-to-face communication since it is one of our body language that increases data information during the communication. In recent surveys, some of the existing methods extracting features from facial images as the regions of interest. Such regions cover eyes and nose, eyes with eyebrows, mouth, etc. Then global features are extracted from those regions afterwards. This feature extraction method can outperform if some irrelevant features are eliminated. Moreover, this causes lower time consumption in the process of normalization and recognition. In this paper, there are two main parts: locating the points in face region to form graph-based features and training the neural networks to recognize the emotion from the corresponding feature vector. For the first phase, fourteen points are manually located to create graph with edges connecting among such points. Subsequently, the Euclidean distances from those edges are calculated and defined as features for training in the next phase. The next phase is using Multilayer-perceptrons (MLPs), a kind of Artificial Neural Networks (ANN), with back-propagation learning algorithm to recognize six basic emotions. In order to evaluate the performance, the proposed systems are applied to Cohn-Kanade AU-Coded facial expression database and perform 95.24% accuracy which is higher than the existing method.

Posture recognition invariant to background, cloth textures, body size, and camera distance using morphological geometry

The human posture estimation in surveillance caring application can improves the people everyday life, In this paper, we propose a method that is invariant to background, distance of camera location, size and cloths of people in the frames. A silhouette is projected to the horizontal and vertical histograms for features extraction. The important features are based on the length and width of body parts of human. The proposed features are more suitable for classifying human posture into four main categories such as standing, lying, sitting, and bending, obviously appeared with the high percentage of recognition when compared with the traditional features in the ANFIS model. The increase of accuracy comes from the robustness of various environments such as the complicated posture of a changed body position and camera distance.

Usability Comparisons of Seven Main Functions for Automated Teller Machine (ATM) Banking Service of Five Banks in Thailand

The objective of this research was to compare the seven main functions of ATM banking services from five banks in Thailand. The selection of the five ATM banks was based on the fact that they contained different hierarchical menu structures. In the research, four groups each with 200 participants were separated into two parts. The first group of participants was required to complete a questionnaire in order to identify the seven main tasks of ATM banking, whilst the second group was required to perform the experiment on the ATM simulator. The second group was subdivided into four groups; students, employees, government and state enterprises officers and agriculturists. To compare seven major functions, a simulator of each of the five banks’ ATM machines was developed and then tested in the laboratory environment. Usability was evaluated in terms of effectiveness, efficiency, satisfaction and the percentage difference. The results suggest that different menu structures will affect the usability of ATM banking. Moreover, the different types of user provide a different score based on usability measurement. Only one bank received the highest score on most of the usability criteria for all the different user groups.

Locating essential facial features using neural visual model

Facial feature detection plays an important role in applications such as human computer interaction, video surveillance, face detection and face recognition. We propose a facial feature detection algorithm for all types of face images in the presence of several image conditions. There are two main step: the facial feature extraction from original face image, and the coverage of the features by rectangular blocks. A neural visual model (NVM) is used to recognize all possibilities of facial feature positions for the first step. Input parameters are obtained from the face characteristics and the positions of facial features not including any intensity information. For the better results, some incorrect decisions of facial feature positions are improved by image processing technique called dilation. Our algorithm is successfully tested with various types of faces which are color images, gray images, binary images, wearing the sunglasses, wearing the scarf, lighting effect, noise and blurring images, color and sketch images from animated cartoon.

A New Design of ATM Interface for Banking Services in Thailand

A new design of ATM interface for banking services in Thailand with a new hierarchical menu structure based on the seven most frequently used tasks is presented. A total, of 105 participants from five different work occupations were used to test the design. Participants were asked to use the new design and a well known existing design. Simulators of the new interface were adapted for testing in a laboratory environment. The HCI principles were considered for designing and testing the new interface using usability criteria as the evaluation. The experimental results showed that a new ATM design reduced the error rate as well as increased effectiveness, efficiency and satisfaction.

A clustering algorithm for stream data with LDA-based unsupervised localized dimension reduction

We present an algorithm for clustering high dimensional streaming data. The algorithm incorporates dimension reduction into the stream clustering framework. When a new datum arrives, the algorithm performs dimension reduction to find a local projected subspace using unsupervised LDA (Linear Discriminant Analysis)-based method. The obtained local subspace would maximally separate the nearby micro-clusters with respect to the incoming point. Then, the incoming point is assigned to a micro-cluster in the projected space, rather than in the full dimensional space. The experimental results show that the proposed algorithm outperforms its counterpart streaming clustering algorithms. Moreover, when compared with traditional clustering algorithms which require the whole data set, the proposed algorithms shows comparable clustering performances with much less computation time for large data sets.