Novian Habibie

Design of intelligent k-means based on spark for big data clustering

The growth of data has bring us to the big data generation where the amount of data cannot be computed using conventional environment. There are a lot of computational environment that had been developed to compute big data, one of them is Hadoop that has Distributed File System and MapReduce framework. Spark is newly framework that can be combined with Hadoop and run on top of it. In this paper, we design intelligent k-means based on Spark for big data clustering. Our design is using batch of data instead using original Resilient Distributed Dataset (RDD). We compare our design with the implementation that using original RDD of data. Result of experiment shows that implementation using batch of data is faster than the implementation using original RDD.

Structured support vector machine learning of conditional random fields

This research aims to improve the capability of semantic segmentation through data perspective. This research proposed a parameterized Conditional Random Fields model and learns the model by using Structured Support Vector Machine (SSVM). The SSVM utilizes Hamming loss function for optimizing 1-slack Margin Rescaling formulation. The joint feature vector is derived from energy potentials. Variation of image size produces some missing values in the joint feature vector. This research shows that a zero padding can resolve the missing values. The experiment result shows that prediction with parameterized CRF yields 75.867% global accuracy (GA) and 22.1410 % averaged class accuracy (CA).

Vehicle traffic monitoring using single camera and embedded systems

Traffic congestion is a problem that often occurs in the big cities in Indonesia. It is caused by very rapid increase of vehicle. The offered solution is to monitor the traffic situation automatically. We implemented the method of detecting vehicle during night in four single board computers (SBC) that are: Raspberry Pi B+, Beagleboard Xm, Raspberry Pi 2 and Odroid XU4. Perfomance of Odroid XU4 exceed other single board computers in which the maximum fps obtained 30 frame per second(fps) and the maximum accuracy of vehicle detection reached 98 percent.

Sensor node for data sampling and correlation analysis of CO 2 concentration with air humidity, temperature, and light intensity

Carbon Dioxide gas (CO2) gas contained in our air which has many roles in environment, but in a huge amount it became dangerous. To encounter that, the system for CO2 monitoring is needed. One of the most effective way is using Wireless Sensor Network (WSN). This system capable to monitor concentration of CO2 and another variable using sensor nodes. But not all of that parameters is correlated to concentration of CO2. To make monitoring system runs efficiently, correlation analysis between variables is needed. This research conduct a correlation analysis between concentration of CO2 and humidity, temperature and light intensity from data collected by our own-made digital sensor node. Data gathered for seven days in one location with a fluctuate environment condition. Correlation calculated with Spearmans rho method. The result is CO2 and air humidity have a strong positive correlation with air humidity (0.726), weak negative correlation with light intensity (−0.319), and no correlation with air temperature (−0.008).

Comparative study of lightweight secure multiroute communication system in low cost wireless sensor network for CO 2 monitoring

Wireless Sensor Network (WSN) is a system used to conduct a remote monitoring in a wide monitoring area. It has a sensor node — a sampling point — which communicate each other to passing their data to central node for recapitulation or transmit it to data center. Because of that, communication system is a crucial thing for WSN. However, WSN may be deployed in a environment that far from ideal condition. Placed in an unattended area with far distance between nodes, WSN is very vulnerable with security threats. To overcome that, the good combination between communication protocol and encryption algorithm for WSN is needed to gather an accurate and representative data with high transmission speed. This research focused on finding those combination for our own-made low-cost sensor node for CO2 monitoring. In this research, two routing protocols (AODV and TARP) and several encryption algorithms (AES, ChaCha, and Speck) tested to determine which combination is give the best result. As the result, combination between routing protocol AODV and encryption algorithm Speck give the best result in the term of performance.

Big sensor-generated data streaming using Kafka and Impala for data storage in Wireless Sensor Network for CO 2 monitoring

Wireless Sensor Network (WSN) is a system that have a capability to conduct data acquisition and monitoring in a wide sampling area for a long time. However, because of its big-scale monitoring, amount of data accumulated from WSN is very huge. Conventional database system may not be able to handle its big amount of data. To overcome that, big data approach is used for an alternative data storage system and data analysis process. This research developed a WSN system for CO2 monitoring using Kafka and Impala to distribute a huge amount of data. Sensor nodes gather data and accumulated in temporary storage then streamed via Kafka platform to be stored into Impala database. System tested with data gathered from our-own made sensor nodes and give a good performance.

Evolutionary segment selection for higher-order conditional random fields in semantic image segmentation

One promising approach for pixel-wise semantic segmentation is based on higher-order Conditional Random Fields (CRFs). We aim to selectively choose segments for the higher-order CRFs in semantic segmentation. To this end, we formulate the selection as an optimization problem. We propose three optimization criteria in relation to the selected segments, namely: a) averaged goodness, b) coverage area and c) non-overlapped area. Essentially, we desire to have best segments with maximum coverage area and maximum non-overlapped area. We apply two evolutionary optimization algorithms, namely: the genetic algorithm (GA) and the particle swarm optimization (PSO). The goodness of segments is estimated using the Latent Dirichlet Allocation approach. Experiment results show that semantic segmentation with GA-or-PSO-selected segments yields competitive semantic segmentation accuracy in comparison to that of naively using all segments. Moreover, the fewer number of segments used in semantic segmentation speeds up its computation time up to six times faster.

Learning semantic segmentation score in weakly supervised convolutional neural network

Semantic segmentation is an image labeling process for each pixels according to defined objects class and its presence in an image. Labeling process consists of recognizing, detecting location and labeling pixels that defines the object in the image. Annotation result of semantic segmentation needs ground truth to verify accuracy of score prediction. Therefore, this research propose a model to predict score of annotation accuracy. By casting the problem into constraining object boundary recognition, we described the annotation using foreground mask. To extract the feature, we used convolution neural network. We only used CNN trained on a image level annotation. In order to be able to infer the pixel instance, we adapt CNN architecture into weakly supervised learning. Experiments were conducted by finetuning Convolution Neural Network for object recognition using weakly supervised architecture for multilabel classification. In this paper we proposed to score semantic segmentation based on bag level information without the availability of pixel level annotation.