Segmentasi Gigi pada Dental Panoramic Radiograph untuk Identifikasi Manusia

<p><em>Abstrak</em> – <strong>Indonesia memiliki tingkat kerawanan tinggi terhadap bencana alam dan kecelakaan yang mengakibatkan terjadinya korban massal. Banyak cara untuk mengidentifikasi korban, salah satunya menggunakan citra gigi. Gigi merupakan bagian dari tubuh yang lebih tahan lama karena struktur gigi yang padat dan kuat. </strong><strong>Identifikasi menggunakan sarana gigi dapat dilakukan dengan cara membandingkan data gigi yang telah diperoleh dari pemeriksaan gigi jenazah yang tidak dikenal (data <em>postmortem</em>) dengan data gigi yang sebelumnya pernah dibuat (data <em>antemortem</em>). Terdapat beberapa tahapan dalam melakukan identifikasi korban menggunakan citra gigi. Tahapan yang dilakukan oleh peneliti adalah tahap segmentasi gigi. Pertama, citra dilakukan <em>cropping</em> hingga mendapatkan dimensi berukuran 1564×589 piksel serta perbaikan citra menggunakan <em>Histogram Equalization</em>. Selanjutnya dilakukan pemisahan citra gigi menggunakan metode <em>Integral Projection</em> dilengkapi penggunaan <em>Spline Interpolation</em> untuk menggambar garis pemisah antara rahang atas-bawah serta gigi tunggal. Tiap citra memiliki nilai n-blok kolom yang berbeda sehingga dibutuhkan parameter sebesar 3 hingga 30 n-blok kolom untuk membentuk garis pemisah rahang atas-bawah. Citra gigi berjenis <em>Dental Panoramic Radiograph</em>. Hasil evaluasi kesalahan<em> </em>terkecil saat melakukan pemisahan rahang atas-bawah menggunakan <em>Horizontal Integral Projection</em> sebesar 56.8% dengan nilai n-blok kolom adalah 8 dan saat<em> </em>segmentasi gigi pada tahap <em>Vertical Integral Projection</em> sebesar 38.27% dengan nilai <em>average filter</em> adalah<em> </em>17.</strong></p><p><em>Abstract</em> – <strong>Indonesia has a high level of vulnerability to natural disasters and accidents that result in mass casualties. There are many ways to identify victims, especially by using dental images. The teeth are part of the body that are more durable because of the solid and strong tooth structure. Identification using dental images can be done by comparing dental data that has been obtained from unknown victim dental examination (postmortem data) with dental data previously made (antemortem data). There are several stages in identifying victims using dental images and researcher worked on tooth segmentation stage. First, the image need to cropped up to get dimensions size of 1564</strong><strong>×</strong><strong>589 pixels and improved contrast using Histogram Equalization method. Then, tooth separation is performed using Integral Projection method which is equipped with the use of Spline Interpolation to draw the separator line between the upper-lower jaws and single tooth. Each image has a different n-block column value, so researcher selected range number of n-block column is between 3-30. In this reseach, dental panoramic radiographs are used. The smallest error rate in the images is found when performing an Integral Projection to separate upper and lower jaws by 56.8% with n-block column value = 8 and when separating each tooth from the image by 38.27% with average filter value = 17.</strong></p><p><strong><em>Keywords</em></strong> – <em>Antemortem, Postmortem, Segmentation, Histogram Equalization, Integral Projection, Spline Interpolation</em></p>

SEISMIC PERFORMANCE OF SCALED IBS BLOCK COLUMN FOR STATIC NONLINEAR MONOTONIC PUSHOVER EXPERIMENTAL ANALYSIS

This paper presents the seismic performance of the down scaled 1:5 model IBS block column with non-linear static analysis. The aim of this research is to access the ultimate capacity and structural behaviour of the IBS block column. This paper demonstrates the theoretical prediction of the full-scale prototype strength based on scaling factors at non-linear state. Besides, this research investigates the ultimate shear capacity, stiffness, bolt strength, inter-storey drift and block separation for prediction of seismic performance levels. Concrete material properties, mix specification and steel reinforcement detailing for scaled model are tabulated in this paper. The methodology of this research begins with full scale prototype design, scaling to the small model and followed by the scaled model fabrication. Theoretical lateral load prediction associated with scaling factors are also performed. The experiment test was carried out on the assembled scaled 1:5 IBS block column with proper displacement measuring equipment on test rig and graphical capture tools. The data of roof top displacement with base shear capacity, inter-storey drift and gap separations were tabulated for discussions. The tested ultimate roof top displacement was 128 mm with 3.1 kN base shear. The calculated elastic stiffness of the IBS block column was 0.137 kN/mm, followed by yielding stiffness of 0.033 kN/mm and 0.014 kN/mm plastic stiffness. The significant inter-storey drift was due to cracking and crushing of column blocks edges. The measured maximum separation gap was 24.4 mm located at 340 mm height due to the rocking of the column. Based on seismic performance levels indicator from FEMA 273 & 356, the column was in the state of immediate occupancy with 21 mm roof top displacement and 1.7 kN base shear. The life safety is limited at 65.27 mm roof top displacement with 2.4 kN of base shear. All scaled down data was then reverted to full scale prototype capacity according with the respective scaling factors. It concluded that the IBS blockwork column is capable of resisting the seismic event without falling of the blocks that endanger the occupant life at the maximum credible earthquakes of 1.3 g horizontal spectral acceleration equivalent to X+ Mercalli’s scale.

Application of the partitioning method to specific Toeplitz matrices

We propose an adaptation of the partitioning method for determination of theMoore-Penrose inverse of a matrix augmented by a block-column matrix. A simplified implementation of the partitioning method on specific Toeplitz matrices is obtained. The idea for observing this type of Toeplitz matrices lies in the fact that they appear in the linear motion blur models in which blurring matrices (representing the convolution kernels) are known in advance. The advantage of the introduced method is a significant reduction in the computational time required to calculate the Moore-Penrose inverse of specific Toeplitz matrices of an arbitrary size. The method is implemented in MATLAB, and illustrative examples are presented.

Corrosive Wear Behavior of Several Materials Coupled with Si3N4 in the Molten Zinc

In continuous hot dip galvanizing, serving parts working in the molten zinc, such as sink roll, sleeves, etc. tend towards degradation and failure due to corrosion and wear. In this paper, corrosive wear performance of several materials, such as boronized，H13，Co-based alloy and Fe-based alloy coupled with Si3N4 is evaluated with the aid of a self-made test machine of block-column sliding wear. Moreover, the corrosive wear mechanism is analyzed. It has been found that these materials suffered not only considerable wear, but also the corrosion of molten zinc. For boronized layer, the wear is the main reason to cause the failure of work parts under the act of the corrosive wear. For H13, the interaction between corrosion and wear plays a very important role on its failure. For those alloys with bad corrosion resistance such as Co-based alloy and Fe-based alloy, both corrosion of solid solution and wear of intermetallic compound act on their failure at the same time. However, the rate of wear shows no obvious relation to the hardness although wear performance is important. The failure greatly lies on the ratio of the hardness of corrosion products to matrix.