KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Settings

Author: search.filters.author.Walairacht, Somsak

Search Results

Now showing 1 - 5 of 5
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Hand gesture recognition for Lao alphabet sign language using HOG and correlation
    (2017-11-03)
    Sombandith, Vimonhak
    ;
    Walairacht, Aranya
    ;
    Walairacht, Somsak
    Hand gesture is one of a powerful means of communication among human. Sign language is an essential and natural expressive mean of communication especially for the deaf people. The article proposes a technique for the recognition of Lao alphabet sign language. The technique of image processing, that is Histogram of Oriented Gradients (HOG), is applied in order to extract characteristics of the hand images performing individual alphabet of Lao sign language. The extracted features are then sent to the template matching process. The similarity between the extracted features and the prototype features are measured by using correlation technique. The totals of 54 Lao alphabets are used in the experiments. Four subjects are asked to perform each alphabet of Lao sign language in which each subject had performed totally 540 gestures. The recognition rate of the proposed technique at about 79 % is achieved.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Recognition of Lao sentence sign language using kinect sensor
    (2017-11-03)
    Sombandith, Vimonhak
    ;
    Walairacht, Aranya
    ;
    Walairacht, Somsak
    This paper presents a technique for recognition Lao sentence sign language by using a Kinect sensor. By using Kinect sensor, feature joint positions of human body can be tracked. After that the joint angles between each pair of joint vectors of adjacent joint positions are measured. When the user perform a series of posture representing a simple sentence of Lao sign language, the system records joint angles that is significantly changed from one posture to another posture. Feature joint angles stored in the database is used to recognition the testing sentence's posture. The recognition rate of about 75% is achieved from the experiments performed 10 simple sentences of Lao sign language.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    PCA in wavelet domain for face recognition
    (2006-11-17)
    Puyati, Wayo
    ;
    Walairacht, Somsak
    ;
    Walairacht, Aranya
    In this paper, the preprocessing process aimed to reduce size of input image by using wavelet transform before transformed image is sent to the process of PCA for recognition. We used ORL Face Databases from AT&T Laboratories Cambridge in the experiments. The results show that the 4<sup>th</sup> Order Symlets level 2 and level 3 improve the accuracy rate of recognition when compare among Haar wavelets, the 4<sup>th</sup> Order Daubechies wavelets, and Biorthogonal wavelets (orthogonal 6.8). In the case of overall processing time for training, the length of filter of wavelet is directly effect the time consuming. Since LL subband of wavelet decomposition becomes the input for PCA, the memory usage can be greatly reduced.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Adaptive spam mail filtering using genetic algorithm
    (2006-11-17)
    Sanpakdee, Usarat
    ;
    Walairacht, Aranya
    ;
    Walairacht, Somsak
    In this paper, we propose a mechanism for filtering incoming spam mails by generating spam mail prototypes using genetic algorithm. Firstly, words from e-mails are extracted and are categorized by their relating meaning into 7 groups. Then, we compose a string of chromosome having 7 genes, i.e., groups of words. Each gene, represented words in each group, is encoded into binary value. The genetic algorithm and its operations are applied to create varieties of spam mail prototypes which inherit from old spam mails. It saves time for preparing training sets and need no large training set for learning like other methods. The spam mail prototypes are the result of this learning mechanism. The experimental results show that the proposed system has efficiency. When testing with both spams and hams, the accuracy is about 85% in average.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Playing Rubik's cube in mixed reality
    (2003-01-01)
    Sato, Makoto
    ;
    Koike, Yasuharu
    ;
    Walairacht, Somsak
    This paper describes a prototype system for virtual objects manipulation in mixed reality environment. SPIDAR-8, the multi-fingers string-based haptic interface device, is used to display force feedback at four fingertips on each hand of the user, i.e., thumb, index, middle, and ring finger. Fingertip positions measured by the haptic interface device are also used to model the 3D virtual hands. Then image sequences of real hands of the user taken by video camera are superimposed on the silhouette of the virtual hands. By employing the technique of sensors fusion and the successive adjustment algorithm, correct geometric registration of image of real hands and the virtual objects can be achieved. The user can see his or her real hands manipulating the virtual objects in the simulated virtual world and perceive force feedback in the same way as manipulating real objects. The system is aimed to enhance the sensation of immerse into the virtual world. © 2003 by Springer Science+Business Media New York.