Search

Heart disease is a heart condition that sometimes causes a person to die suddenly. One indication is a rhythm disorder known as arrhythmia. Multi-class Arrhythmia Detection has followed: QRS complex detection procedure and arrhythmia classification based on the QRS complex morphology. We proposed an edge device that detects QRS complexes based on variance analysis (QVAT) and the arrhythmia classification based on the QRS complex spectrogram. The classifier uses two-dimensional convolutional neural network (2D CNN) deep learning. We use a single board computer and neural network compute stick to implement the edge device. The outcomes are a prototype device cardiologists use as a supporting tool for analysing ECG signals, and patients can also use it for self-tests to figure out their heart health. To evaluate the performance of our edge device, we tested using the MIT-BIH database because other methods also use the data. The QVAT sensitivity and predictive positive are 99.81% and 99.90%, respectively. Our classifier's accuracy, sensitivity, predictive positive, specificity, and F1-score are 99.82%, 99.55%, 99.55%, 99.89%, and 99.55%, respectively. The experiment result of arrhythmia classification shows that our method outperforms the others. Still, for r-peak detection, the QVAT implemented in an edge device is comparable to the other methods. In future work, we can improve the performance of r-peak detection using the double-check algorithm in QVAT and cross-check the QRS complex detection by adding 1 class to the classifier, namely the non-QRS class.

A freehand three-dimensional (3D) ultrasound system is a method of acquiring images using a 3D ultrasound probe or conventional two-dimensional (2D) ultrasound probe to give a 3D visualization of an object inside the body. Ultrasounds are used extensively in clinical applications since they are advantageous in that they do not bring dangerous radiation effects and have a low cost. However, a probe calibration method is needed to transform the coordinate position into a 3D visualization display, especially for image-guided intervention. The current ultrasound probe calibration system usually uses the numerical regression method for the N-wire phantom, which has problems in accuracy and reliability due to nonlinear point scattered ultrasound image data. Hence, a method for ultrasound probe positional calibration of single-wire phantom using the Levenberg-Marquardt algorithm (LMA) was proposed to overcome this weakness. This experiment consisted of an optical tracking system setup, a 2D ultrasound probe with marker, an ultrasound machine, and a single-wire object in a water container equipped with a marker. The position and orientation of the marker in a 2D ultrasound probe and the marker in the water container were tracked using the optical tracking system. A 2D ultrasound probe was equipped with a marker connected wirelessly using an optical tracking system to capture the single-wire object. The resulting sequences of 2D ultrasound images were reconstructed and visualized into 3D ultrasound images using three transformations, ultrasound beam to ultrasound probe’s marker, single-wire phantom position to container’s marker, and the 3D visualization transformation. The LMA was used to determine the best optimization parameters for determining the exact position and representing that 3D visualization. The experiment result showed that the lowest mean square error (MSE), rotation error, and translation error were 0.45 mm, 0.25°, and 0.3828 mm, respectively.

Abstract Computer‐aided craniofacial reconstruction (CFR) is a process that aims to estimate facial impressions based on skull remains. It mimics the conventional method with a conceptual model‐based framework. The existing problems in CFR are that landmark annotation is expert‐dependent, landmark processing in the 3D domain has volumetric challenges, and a method based on a population's morphological characteristics (templates). A framework with three stages is proposed: Building a craniofacial model, automatic landmark detection, and surface deformation. Machine learning is deployed to draw local surface correlations as landmarks and automatically detects their position. The local surface context is extracted using the Surface Curvature Feature (SCF) as a 3D descriptor. Using a cluster‐based filter, the average distance (to the ground truth) of the top 20 points is 0.0326 units. Cluster‐based filters are better than mass‐radius‐based filters and consistently give better pinpoint accuracy, especially in multi‐cluster cases. Training data consists of 140,000 SCF for ten landmark classes. The third stage, surface deformation, fits the facial template to the cranial based on the corresponding facial‐cranial landmarks. Five experts from the Anthropology department stated that of the reconstruction results, 91.5% could retain the template details and are accepted as the natural shape of the human face.

Person re-identification is one of the problems in the computer vision field that aims to retrieve similar human images in some image collections (or galleries). It is very useful for people searching or tracking in a closed environment (like a mall or building). One of the highlighted things on person re-identification problems is that the model is usually designed only for performance instead of performance and computing power consideration, which is applicable for devices with limited computing power. In this paper, we proposed a lightweight residual network with pyramid attention for person re-identification problems. The lightweight residual network adopted from the residual network (ResNet) model used for CIFAR dataset experiments consists of not more than two million parameters. An additional pyramid features extraction network and attention module are added to the network to improve the classifier's performance. We use CPFE (Context-aware Pyramid Features Extraction) network that utilizes atrous convolution with different dilation rates to extract the pyramid features. In addition, two different attention networks are used for the classifier: channel-wise and spatial-based attention networks. The proposed classifier is tested using widely use Market-1501 and DukeMTMC-reID person re-identification datasets. Experiments on Market-1501 and DukeMTMC-reID datasets show that our proposed classifier can perform well and outperform the classifier without CPFE and attention networks. Further investigation and ablation study shows that our proposed classifier has higher information density compared with other person re-identification methods.

Cranial anthropometric reference points (landmarks) play an important role in craniofacial reconstruction and identification. Knowledge to detect the position of landmarks is critical. This work aims to locate landmarks automatically. Landmarks positioning using Surface Curvature Feature (SCF) is inspired by conventional methods of finding landmarks based on morphometrical features. Each cranial landmark has a unique shape. With the appropriate 3D descriptors, the computer can draw associations between shapes and landmarks using machine learning. The challenge in classification and detection in three-dimensional space is to determine the model and data representation. Using three-dimensional raw data in machine learning is a serious volumetric issue. This work uses the Surface Curvature Feature as a three-dimensional descriptor. It extracts the local surface curvature shape into a projection sequential value (depth). A machine learning method is developed to determine the position of landmarks based on local surface shape characteristics. Classification is carried out from the top-n prediction probabilities for each landmark class, from a set of predictions, then filtered to get pinpoint accuracy. The landmark prediction points are hypothetically clustered in a particular area, so a cluster-based filter is appropriate to isolate them. The learning model successfully detected the landmarks, with the average distance between the prediction points and the ground truth being 0.0326 normalized units. The cluster-based filter is implemented to increase accuracy compared to the ground truth. Thus, SCF is suitable as a 3D descriptor of cranial landmarks.

When the blood flow to the arteries in brain is blocked, its known as Ischemic stroke or blockage stroke. Ischemic stroke can occur due to the formation of blood clots in other parts of the body. Plaque buildup in arteries, on the other hand, can cause blockages because if it ruptures, it can form blood clots. The b-1000 Diffusion Weighted (DW) Magnetic Resonance Imaging (MRI) image was used in a general examination to obtain an image of the part of the brain that had a stroke. In this study, classifications used several variations of layer convolution to obtain high accuracy and high computational consumption using b-1000 Diffusion Weighted (DW) MR in ischemic stroke types: acute, sub-acute and chronic. Ischemic stroke was classified using five variants of the Convolutional Neural Network (CNN) architectural design, i.e., CNN1–CNN5. The test results show that the CNN5 architectural design provides the best ischemic stroke classification compared to other architectural designs tested, with an accuracy of 99.861%, precision 99.862%, recall 99.861, and F1-score 99.861%.

Gangguan tidur adalah terganggunya aliran pernapasan (kolaps) yang berulang saat tidur. Gangguan tidur umumnya tidak terdiagnosis dan tidak tertangani dengan baik sehingga terjadi interupsi berulang pada saluran pernapasan. Kondisi tersebut berpotensi menyebabkan penderita mengalami gangguan tidur yang serius. Salah satu alat bantu yang digunakan untuk menganalisis adanya gangguan tidur pada ritme jantung adalah melalui rekaman electrocardiogram (ECG). ECG merupakan suatu grafik yang merepresentasikan aktivitas listrik pada jantung berupa gelombang P, gelombang QRS, dan gelombang T. Banyaknya sensor untuk merekam ECG menyebabkan ketidaknyamanan pasien saat perekaman data, menyebabkan kompleksitas rekaman, lambatnya komputasi, kesalahan dalam interpretasi, dan kehilangan informasi klinis. Oleh karena itu, dibutuhkan suatu sistem peringatan dini sebagai alat bantu medis yang mampu mendiagnosis rekaman ECG menggunakan satu sadapan. Sistem didesain menjadi lima tahapan, yaitu akuisisi rekaman ECG, praproses, ekstraksi fitur, seleksi fitur, dan proses klasifikasi. Fitur set yang diperoleh dari rekaman ECG berupa variabel-variabel yang diperoleh dari domain waktu, domain frekuensi, dan analisis nonlinear. Pengujian menunjukkan metode AdaBoost memiliki nilai kinerja paling tinggi dibandingkan dengan SVM, k-NN, dan NN. Metode AdaBoost secara berurutan memiliki kinerja classification accuracy, F1 score, precision, recall, sensitivity, dan specificity sebesar 90,1%, 90,1%, 90,3%, 90,1%, 86,9%, dan 93,3%.

Detection of vascular areas (blood vessels) using B-Mode ultrasound images is needed for automatic applications such as registration and navigation in medical operations. This study developed the detection of the carotid artery area using Convolution Neural Network Single Shot Network Multibox Detector (SSD) to determine the bounding box ROI of the carotid artery area in B-mode ultrasound images. The data used are B-Mode ultrasound images on the neck that contain the carotid artery area (primary data). SSD method result is 95% of accuracy which is higher than the Hough transformation method, Ellipse method, and Faster RCNN in detecting carotid artery area in the B-Mode ultrasound image. The use of image enhancement with Gaussian filter, histogram equalization, and Median filters in this method can increase detection accuracy. The best process time of the proposed method is 2.09 seconds so that it can be applied in a real-time system.

Teknologi komputer telah banyak digunakan di tidak hanya untuk tujuan penelitian dan pendidikan saja tetapi juga di dunia hiburan. Salah satu hiburan berbasis komputer adalah permainan komputer dan animasi. Salah satu komponen pendukungnya adalah machinima. Machima adalah teknologi yang akan menempatkan sebuah komponen sinematik dalam dunia virtual. Salah satu komponen yang dapat dikontrol adalah penempatan posisi kamera. Seorang sutradara bisa dibedakan gaya penyutradaraannya, salah satunya berdasarkan penempatan posisi kamera. Dengan menerapkan suatu gaya penyutradaraan pada sebuah permainan atau animasi bisa mendapatkan suasana yang berbeda. Penelitian ini akan mencoba melakukan profiling terhadap gaya seorang sutradara berdasarkan kebiasaan penempatan posisi kamera. Pendekatan yang dilakukan berbasis logika fuzzy. Penelitian ini akan menggunakan 19 variabel input yang berasal dari hasil ektraksi data simulasi dan 5 variabel output untuk melakukan profiling terhadap dua buah gaya sutradara yang berbeda dengan pendekatan logika fuzzy. Akan dihasilkan diagram area dan histogram sehingga mempermudah pembacaan dalam membedakan gaya sutradara dan berhasil dibedakan berdasarkan modus hasil analisa terhadap diagram histogram.

Penempatan posisi kamera secara manual membutuhkan proses dan perhitungan yang berulang-ulang untuk sebuah adegan, tentu saja hal ini akan membutuhkan banyak waktu dan biaya. Oleh karena itu dibuat sebuah sistem penempatan posisi kamera secara otomatis dalam lingkungan virtual yang sesuai dengan kaidah sinematografi. Sistem akan menangani pengaturan kamera untuk mendapat sudut, jarak dan ketinggian yang cocok untuk setiap adegan. Penelitian ini akan menggunakan simulasi permainan komputer dengan beberapa aksi dengan dua macam gaya penempatan posisi kamera yaitu kamera statis dan kamera dinamis. Metode yang digunakan adalah logika fuzzy dengan metode mamdani. Terdapat 3 tahapan yang dilakukan yaitu tahap masukan dari pengguna berupa kombinasi penekanan tombol untuk melakukan sebuah aksi. Tahap kedua adalah sistem inferensi fuzzy yang akan mengolah input menjadi output. Tahap terakhir adalah perekaman aksi berdasarkan posisi kamera yang telah ditentukan. Selain itu dalam simulasi permainan ini akan dipasang modul event selector untuk validasi aksi dan modul state director untuk operator transisi kamera virtual. Uji coba dilakukan dengan cara merekam 7 macam aksi dan akan divalidasi berdasarkan profiling dan responden. Dari uji coba didapatkan nilai 3.66 untuk kamera statis dan 4.08 untuk kamera dinamis. Hal ini menunjukkan bahwa sistem yang dibangun mampu melakukan penempatan posisi kamera sesuai kaidah sinematografi.

Femoral-tibial alignment is a prominent risk factor for Knee Osteoarthritis (KOA) incidence and progression. One way of assessing alignment is by determining the Femoral-Tibial Angle (FTA). Several studies have investigated FTA determination; however, methods of assessment of FTA still present challenges. This paper introduces a new method for semi-automatic measurement of FTA as part of KOA research. Our novel approach combines preprocessing of X-ray images and the use of Active Shape Model (ASM) as the femoral and tibial segmentation method, followed by a thinning process. The result of the thinning process is used to predict FTA automatically by measuring the angle between the intersection of the two vectors of branching points on the femoral and tibial areas. The proposed method is trained on 10 x-ray images and tested on 50 different x-ray images of the Osteoarthritis Initiative (OAI) dataset. The outcomes of this approach were compared with manually obtained FTA measurements from the OAI dataset as the ground truth. Based on experiments, the difference in measurement results between the FTA of the OAI and the FTA obtained using our method is quite small, i.e., below 0.81° for the right FTA and below 0.77° for the left FTA with minimal average errors. This result indicates that this method is clinically suitable for semi-automatic measurement of the FTA.

Teknologi komputer saat ini telah banyak digunakan dalam pengembangan animasi atau permainan komputer. Salah satu teknologi itu adalah machinima yaitu suatu sistem yang menggunakan teknologi mesin grafik 3D untuk menghasilkan produk sinematik secara real time. Dalam proses pembuatan produk sinematik itu penempatan posisi kamera sangat memegang peranan penting. Penempatan posisi kamera ini tentu harus sesuai dengan kaidah-kaidah sinematografi. Penelitian ini akan mengusulkan sebuah pendekatan agen cerdas dengan multi perilaku untuk menempatkan kamera virtual dalam lingkungan virtual secara otomatis sesuai dengan gaya seorang sutradara. Setiap kamera virtual itu akan memiliki perilaku yang berbeda berdasarkan kaidah sinematografi sehingga memiliki Point of View (POV) yang berbeda. Untuk memberikan perilaku pada kamera virtual akan digunakan pendekatan berbasis logika fuzzy dengan menggunakan metode mamdani. Jumlah variabel masukan yang digunakan sejumlah tiga dan variabel keluaran sejumlah tiga dengan membership function antara tiga sampai lima. Penelitian ini akan menggunakan simulasi permainan komputer dengan tiga kamera virtual dengan perilaku yang berbeda untuk merekam adegan yang sama dan hasilnya akan divalidasi berdasarkan hasil pengamatan dengan komunitas juru foto. Pada akhirnya dapat diambil kesimpulan bahwa pendekatan logika fuzzy dapat digunakan untuk memberikan sebuah perilaku atau gaya sutradara pada kamera virtual. Abstract Computer technology is has been used widely in the development of animation or computer games. One of the technologies is machinima, a system that uses reak time 3D graphics engine technology to produce cinematic products. In the process of develop a cinematic product, camera positioning is a very important component. The camera positioning must be comply with cinematography’s rule. This research will propose an intelligent multi agent behavior to positining a virtual camera in a virtual environment automatically according to the director’s style. Each virtual camera will have a different behavior based on cinematographic rules so that it has a different Point of View (POV). To assign a behavior on the virtual camera will be based on fuzzy logic using the mamdani method. The number of input variables are three and the output variables are three with the number membership functions between three to five. This research will program a computer game simulation with three multi behavior virtual cameras to capture some scene and the results will be validated based on observations with the photographer community. Finally it can be concluded that the fuzzy logic approach can be used to assign some behavior to a virtual camera.

In this paper, we proposed a parallel spatial pyramid CNN classifier for image-based kinship verification problem. Two face images that compared for kinship verification treated as input for each parallel convolutional network of our classifier. Each parallel convolutional network constructed using spatial pyramid CNN classifier. At the end of the convolutional network, we use three fully connected layers to combine each spatial pyramid CNN features and decided the final kinship prediction. We tested the proposed classifier using large-scale kinship verification dataset, called FIW dataset, consists of seven kinship problems from 1,000 families. In our approach, we treated each kinship problem as a binary classification problem with two output. We train our classifier separately for each kinship problem with same training configuration. Overall, our proposed method can achieve an average accuracy of more than 60% and outperform the baseline method.

Osteoarthritis (OA) is a joint disease that affects a large part of the elderly population. One of the OA that is often experienced by patients is knee OA. To determine the development and classification of this disease, a process of segmenting the femur and tibia is needed quickly and accurately. Meanwhile, manual segmentation has several disadvantages including the longer time needed and the difference in the results of reading x-ray images between medical personnel with each other. Therefore, in this paper, an Adaptive Active Shape Model (AASM) is presented for femur and tibia segmentation on knee x-ray images. The purpose of this segmentation is to support the discovery and characterization of imaging biomarkers for the incidence, clinical evaluation, classification, and progression of knee osteoarthritis (OA). This new algorithm is adaptively capable of better segmenting the femur and tibia than the original ASM. In this experiment, 10 images were used as training data to get the mean shape model and 50 images were tested to find out performance of the method implemented. All images are taken randomly from Osteoarthritis Initiative (OAI) dataset. To determinate the accuracy of this segmentation method, calculations have been performed using Hausdorff Distance (HD) and Dice Similarity Coefficient (DSC). In addition, this study have also been compared with previous research (original ASM) and the same data is used. The best average result of the segmentation validation method from 50 test images in the AASM method using HD is 0.2016 for the right tibia femur bone using 43 landmarks and 0.9497 for the DSC. Based on these results, the average increase in accuracy of segmentation validation was 0.29 for HD and 0.33 for DSC. Thus, this method is quite reliable and clinically valuable for monitoring the progression of knee osteoarthritis

MANET is autonomous, self-configured, and applicable to emergency locations such as forest fires, earthquakes, floods, and health monitoring. However, challenges and difficulties faced by the mobile ad-hoc network (MANET) is a dynamically built network system, without the support of infrastructure in communicating between one node and other nodes, and limited energy sources. To overcome MANET problems and to obtain optimal network quality, the selections of routing protocols and quality of service (QoS) are significant in MANET design. This study aims to analyze the performance of routing protocols: dynamic source routing (DSR), ad-hoc on demand distance vector (AODV) and optimized link state routing (OLSR) based on QoS. The analyzed QoS parameters include packet delivery ratio (PDR), packet loss, throughput, and delay. Simulation results using network simulator version based on the number of node densities indicate that OLSR has better performance compared to AODV and DSR regarding PDR, packet loss, throughput, and delay.

Babies are still unable to inform the pain they experience, therefore, babies cry when experiencing pain. With the rapid development of computer vision technologies, in the last few years, many researchers have tried to recognize pain from babies expressions using machine learning and image processing. In this paper, a research using Deep Convolution Neural Network (DCNN) Autoencoder and Long-Short Term Memory (LSTM) Network is conducted to detect cry and pain level from baby facial expression on video. DCNN Autoencoder is used to extract latent features from a single frame of baby face. Sequences of extracted latent features are then fed to LSTM so the pain level and cry can be recognized. Face detection and face landmark detection is also used to frontalize baby facial image before it is processed by DCNN Autoencoder. From the testing on DCNN autoencoder, the result shows that the best architecture used three convolutional layers and three transposed convolutional layers. As for the LSTM classifier, the best model is using four frame sequences

The development of mobile ad-hoc network (MANET) is becoming popular and interesting to research because it has a fast characteristic, cost-effective deployment, able to manage topology change independently, and can be applied to emergency locations such as forest fire detection, military operation, and health monitoring. However, the problems faced by MANET are dynamic network topology changes, limited energy consumption, and without the support of existing infrastructure. In order to overcome dynamic topology changes and to obtain reliable network quality, then routing protocol selection is critical in designing MANET. This study aims to analyze the performance of AODV, DSR, DSDV, and OLSR routing protocols based on the quality of service (QoS). Scenarios based on the number of vertices, packet size, a broad area of simulation, length of simulation, simulation speed, mobility model, and propagation model. The simulation has been done to produce four graphs, each of which describes the PDR, throughput, packet loss, and delay. The simulation results show that OLSR performs better than DSR, AODV, and DSDV in terms PDR, throughput, packet loss, and delay. OLSR average value in PDR by 39.997%, throughput by 417.383 Kbps, packet loss by 60.003%, and delay of 15.52 milliseconds.

Along with the rapid growing world of technology nowadays, application becomes an important requirement for every human in person. Human becomes dependent on variety of applications, ranging from small scale application for entertaining user to large scale web application for one or multiple businesses work. In this research the application that used is social and online media monitoring at PT. XYZ in Indonesia. Based on this case, a software quality testing is needed so a software can be said worthy and qualified to use, not just limited to mere testing, but also for optimizing the software to create a well maintained application in accordance with software development life cycle. In this paper, WebQEM is used in optimization for web application. There are two kinds of evaluations: basic evaluation and global evaluation. First global evaluation of monitoring social media and online media application gives a score of 70,44%. After evaluation, an improvement is applied to the application according to the criteria from the result of first global evaluation. Second global evaluation gives a value of 77,41%, from the improvement in first global evaluation. This proves that WebQEM method for optimization of monitoring social media and online media application can improve the software quality.

Machinima is a computer imaging technology typically used in games and animation. It prints all movie cast properties into a virtual environment by means of a camera positioning. Since cinematography is complementary to Machinima, it is possible to simulate a director’s style via various camera placements in this environment. In a gaming application, the director’s style is one of the most impressive cinematic factors, where a whole different gaming experience can be obtained using different styles applied to the same scene. This paper describes a system capable of automatically profile a director’s style using fuzzy logic. We employed 19 output variables and 15 other calculated variables from the animation extraction data to profile two different directors’ styles from five scenes. Area plots and histograms were generated, and, by analyzing the histograms, different director’s styles could be subsequently classified.

This research develops the settlement mechanism in the Real Time Gross Settlement using so called clearing house through a serious game method. In this approach banks are represented as nodes that do the settlement process according to the simple rules. Moreover, the graph mining approach is used for predicting the activity networks on those banks. As the result, for constant nodes indicate that the more the activity networks among banks are available, the more the activity networks can be identified. Furthermore, the smaller the differences among the bank health’s level are, the greater the network activities can be identified. This behavior is a consequence of chosen fixed point assumption.

DETERMINING THE ABNORMALITY OF BULL SPERM TAIL MORPHOLOGY USING SUPPORT VECTOR a Stevanus Hardiristanto, b I Ketut Eddy Purnama, cAdhi Dharma Wibawa, dMira Candra Kirana, eBudi Santoso, fMunawir, g Slamet Hartono, h I Nyoman Tirta Ariana, iDian Ratnawati, jLukman Affandhy a,b,c,d,e,fDepartment of Multimedia and Network Engineering, Faculty of Industrial Technology, Institute of Technology Sepuluh Nopember, Surabaya, Indonesia gBalai Pembibitan Ternak Unggul Sapi Bali, Ministry of Agriculture, Republik of Indonesia h Faculty of Animal Science, University of Udayana, Bali, Indonesia i,jLoka Penelitian Sapi Potong Grati, Ministry of Agriculture, Republik of Indonesia E-mail: a hardi@its.ac.id Abstrak Penilaian atas ketidaknormalan spermatozoa bisa dilakukan dari sisi motilitas maupun morfologi (kepala dan ekor). Penelitian ini mengevalusi ketidaknormalan spermatozoa dari sisi morfologi bagian ekor spermatozoa sapi. Data berupa 50 citra mikroskopis spermatozoa yang diperoleh dari Loka Penelitian Sapi Potong Grati, Pasuruan digunakan dalam penelitian ini. Prosedur yang ditetapkan terdiri atas beberapa tahap. Tahap pertama adalah melakukan segmentasi spermatozoa untuk memisahkan spermatozoa dari latar belakang dan memisahkan bagian ekor spermatozoa dari bagian yang lain. Selanjutnya dari hasil segmentasi dicari garis tengah ekor (skeleton) menggunakan metode medial axis transform. Berdasarkan garis tengah yang dihasilkan, dilakukan prosedur ekstraksi fitur menggunakan metode polynomial curve fitting. Kemudian, metode Support Vector Machine (SVM) digunakan untuk menentukan ketidaknormalan bentuk ekor spermatozoa. Untuk pembelajaran digunakan 25 data spermatozoa normal dan 10 data spermatozoa tidak normal. Testing kemudian dilakukan atas 15 data spermatozoa tersisa. Ketelitian SVM dalam menentukan ketidaknormalan bentuk ekor spermatozoa mencapai 73.33%. Dengan demikian ketidaknormalan bentuk ekor spermatozoa dapat ditentukan dengan menggunakan SVM. Kata kunci: Ekor Sperma sapi, Morphology, Polynomial Curve Fitting, SVM. Abstract Determinining the abnormality of spermatozoa can be done by inspecting its motility or morphology (head or tail). This study examined 50 data of sperm microscopic images. The semen was obtained from Loka Penelitian Sapi Potong Grati, Pasuruan. A sequence of procedure consist of several steps were then carried out. The first step was to obtain sperm tails by segmenting the sperms from its background and removing the heads and the necks parts. The skeletons of the tails were then obtained using a method of medial axis transform. The features of the tails were then extracted using polynomial curve fitting. Then, Support Vector Machine (SVM) was used as a classifier. In the training phase, 25 normal sperm and 10 abnormal sperm were utilized. Afterward, the remaining 15 data were used in the testing phase. The accuracy of SVM was 73.33%. Hence, the abnormality of spermatozoa based on the shape of sperm tail can be determined using SVM. Key words: Bull Sperm Tail, Morphology, Polynomial Curve Fitting, SVM

ENHANCEMENT OF 3D SURFACE RECONSTRUCTION OF UNDERWATER CORAL REEF BASE ON SIFT IMAGE MATCHING USING CONTRAST LIMITED ADAPTIVE HISTOGRAM EQUALIZATION AND OUTLIER REMOVAL aPulung Nurtantio Andono, bRicardus Anggi Pramunendar, cCatur Supriyanto, dGuruh Fajar Shidik,e I Ketut Eddy Purnama, fMochamad Hariadi a,b,c,dFaculty of Computer Science, Dian Nuswantoro University Jalan Imam Bonjol, No. 207, Semarang 50131, Indonesia e,fFaculty of Industrial Technology, Dept. of Electrical Engineering, ITS, Surabaya, Indonesia Email: a pulung@research.dinus.ac.id Abstrak Penelitian ini menggambarkan peningkatan kualitas rekonstruksi 3D permukaan terumbu karang bawah laut menggunakan sistem kamera stereo. Algoritma Contrast Limited Adaptive Histogram image Equalization (CLAHE) diusulkan untuk meningkatkan kualitas citra bawah laut tersebut, karena menurunnya kualitas citra bawah laut dapat disebabkan oleh penyerapan dan hamburan sinar matahari. Dalam mengembangkan rekonstruksi 3D permukaan bawah laut, pasangan citra stereo diekstrak secara manual dari rekaman video yang diperoleh, yang kemudian dilakukan proses pencocokan citra stereo menggunakan algoritma SIFT. Kelebihan algoritma SIFT tersebut adalah tahan terhadap perubahan skala, transformasi, dan rotasi dari sepasang citra tersebut. Banyaknya matching point antar 2 citra stereo dijadikan ukuran untuk mengetahui kinerja CLAHE terhadap algoritma SIFT. Hasil penelitian menunjukan bahwa penggunaan CLAHE dan outlier removal mampu meningkatkan jumlah matching point sebesar 56%. Keberhasilan CLAHE tersebut perlu diujikan ke beberapa algoritma matching point yang lain. Perbandingan beberapa algoritma matching point yang menerapkan CLAHE dapat membuktikan bahwa CLAHE sangat cocok dalam meningkatkan kinerja algoritma matching point dan rekonstruksi permukaan 3D citra bawah laut. Kata kunci: Rekonstruksi 3D, Citra Bawah Laut, SIFT, CLAHE. Abstract This research describes an enhancement of 3D Reconstruction coral reef images using stereo camera system. Contrast Limited Adaptive Histogram image Equalization (CLAHE) algorithm was proposed to enhance the image quality in preprocessing area, since the quality of underwater images degrades by the absorption and scattering of light. To develop a 3D-representation of the seafloor, image-pairs were first extracted from the video footage manually, then corresponding points are automatically extracted from the stereo-pairs by SIFT matching algorithm, which is invariant to scale, translation, and rotation. Number of matching points is used to evaluate the performance of SIFT with and without CLAHE. As a result, the promising techniques provides better 3D reconstruction details of coral reef imagesin total, the combination of CLAHE and outlier removal performs the enhancement for 56%. For further, CLAHE need to be performed to other image matching techniques. The comparison of different image matching techniques with and without CLAHE can prove that CLAHE is appropriate as image enhancement method for image matching and 3D surface reconstruction. Key words: 3D Reconstruction, Underwater Image, SIFT, CLAHE.