Indah Soesanti

Identification of malignant masses on digital mammogram images based on texture feature and correlation based feature selection

The most popular techniques in early breast cancer detection is using digital mammogram. However, the challenge lies in early and accurate detection the irregular masses with spiculated margin as the most common abnormality. This paper proposes an image classifier to classify the mammogram images. The abnormality that can be founded in mammogram image is classified into malignant, benign and normal cases. By applying Computer Aided Diagnosis (CAD), totally 12 features comprising of histogram and GLCM as the texture based features are extracted from the mammogram image. Correlation based feature selection (CFS) is used in this paper which reduces 50% of the features. Multilayer perceptron algorithm is applied to mammography classification by using these selected features. The experimental result shows that 40 digital mammograms data taken from private Oncology Clinic Kotabaru Yogyakarta was achieved 91.66% of accuracy. The approach can be beneficial to radiologists for more accurate diagnosis.

Portable smart sorting and grading machine for fruits using computer vision

This paper discusses the development of portable fruit sorting and grading machine based on computer vision for small agro-industries. The mechanical system is designed from low cost material in the form of inclined and segmented plane to substitute the utilization of conveyor belt. In this case, motor servos are used as gate opener and director for the mechanical system. The autonomous system collects video image from a Logitech C920 webcam placed on the top of analysis area, then the image will be analyzed due to the process of computer vision. Firstly, the computer vision algorithm transforms the RGB (Red, Green, and Blue) color space to HSV (Hue, Saturation, and Value) color space of the image to facilitate the processes of color segmentation that are robust to the light intensity fluctuation. To speed up the process, every single frame is classified to 2 ROI (Region of Interest) based on fruit position in queuing and analysis area. Then the system will cluster fruit quality according to the level of maturity and its dimension. In the end, the autonomous system will actuate the servos to move the fruit to a specific bin according to their quality grade. Then the result of fruit analysis data will be displayed on PCs monitor. The system can do the task in 500 ms with precision result.

Power System Stabilizer model based on Fuzzy-PSO for improving power system stability

This paper proposed a Power System Stabilizer model based on Fuzzy-PSO for improving power system stability. Power System Stabilizer (PSS) is a device that can be used to enhance the damping of power system during low frequency oscillations. In multi-machine power systems, the PSS parameter tuning is a complex exercise due to the presence of several poorly damped modes of oscillation. The problem is further being complicated by continuous varied in power system operating conditions. In order to enhance the performance of PSS, the combination of fuzzy logic and particle swarm optimization (PSO) method is used in this study. Simulations were carried out using several fault tests at transmission line on a two-area multi-machine power system. In this work, Delta w PSS and Delta Pa PSS has been used for comparison with the fuzzy-PSO PSS. The result shows that power transfer response using the fuzzy-PSO PSS is more robust than Delta w PSS and Delta Pa PSS, especially for three phase faults and phase to ground faults.

Performance analysis of edge and detailed preserved speckle noise reduction filters for breast ultrasound images

Breast ultrasound is one of the powerful modalities for medical breast lesion imaging. Since ultrasound images are usually corrupted by speckle noise, it is important to perform effective despeckling process. Some filters perform despeckling process smoothly but they do not preserve the edges and details of ultrasound images. Edges and details are important features of lesion classifications. This paper aims at studying speckle noise reduction filters performance due to their ability for preserving edges and details of breast ultrasound images. Adaptive median filter, Frosts filter, detailed preserved anisotropic diffusion filter and Wiener filter in wavelet domain are compared and evaluated on the basis of Peak Signal to Noise Ratio, Mean Squared Error, Average Difference, Mean, and Variance as second-order texture operator.

Analysis of digital mammograms for detection of breast cancer

Digital mammogram has become the most effective technique for early breast cancer detection. The most common abnormality that may indicate breast cancer is masses. The challenge lies in early and accurate detection to overcome the development of breast cancer that affects more and more women throughout the world. Computer Aided Diagnosis (CAD) is used to help the radiologist in interpretation and recognition the pattern of the mammogram abnormality. The main objective of this research is to perform and analyze the contrast enhancement and feature selection method in order to build a CAD to discriminate normal, benign, and malignant. Preprocessing needs to enhance the poor quality of image and remove the artifact caused by preprocessing step. ROI as the suspicious area segmented, and then extracted by texture feature approach. High dimensionality of feature is selected by feature selection technique and would be classified according to their class each other. The digital mammogram images are taken from the Private database of Oncology Clinic Kotabaru Yogyakarta. The dataset consists of 40 mammogram images with 14 benign cases, 6 malignant cases, and 20 normal cases. The proposed method in preprocessing step made the image enhanced and proved by MSE and PSNR value. Histogram and gray level co-occurrence matrix (GLCM) as the texture feature are used to extract the suspicious area. Correlation based feature selection (CFS) is used to select the best feature among 12 extracted features before. Mean, standard deviation, smoothness, angular second moment (ASM), entropy, and correlation are the best feature that guarantee the improvement of classification with less feature dimension. The result shows that the proposed method was achieved the accuracy 96.66%, sensitivity 96.73%, specificity 97.35% and ROC 96.6% It is expected to contribute for helping the radiologist as material consideration in decision-making.

Classification of breast ultrasound images based on texture analysis

Ultrasonography (USG) is a popular imaging modality because of its flexibility, non-invasion, non-ionisation and low cost. A breast ultrasound used to detect and classify abnormalities of the breast mass. However, the diagnosis is very subjective because it depends on the ability of the radiologist. In order to eliminate operator dependency and to improve the diagnostic accuracy, a computerised system is necessary to do the feature extraction and the classification of the breast nodule. This research proposes a classification of breast USG images by using some texture features into two classes. The dataset consists of 57 USG images which grouped into 27 anechoic cases and 30 hypoechoic cases. An initial step of image pre-processing is conducted to enhance the detection capability. Afterwards, followed by some methods of morphological operation, region growing active contour and histogram equalization. The feature extraction method used texture analysis, which is histogram, gray level co-occurrence matrix (GLCM) and fractal Brownian motion (FBM). Finally, Multilayer Perceptron (MLP) classification method is used to classify anechoic nodule from hypoechoic nodule. The result shows that the proposed method achieved the accuracy of 91.23%, sensitivity of 95.83%, specificity of 87.88%, Positive Predictive Value (PPV) of 85.19% and Negative Predictive Value (NPV) of 96.67%. This suggest that the proposed method is excellent in analyzing breast USG images.

Heart sound feature extraction and classification using autoregressive power spectral density (AR-PSD) and statistics features

Heart auscultation is a screening method done by listening using a stethoscope for early diagnosis of heart disease, it is low cost and non-invasive, but it has a limitation of human hearing. This paper presents a noise-robust feature extraction method by combining and selecting a heart sound (HS) feature in time and frequency domain. Wavelet decomposition (WD) is used for noise removal. Features are extracted by AR-PSD and used as inputs for classification. The nine types of abnormal HS that were taken from Michigan Heart Sound Database were classified into nine categories. In this research, statistics features in time and frequency domain are used as additional features. Correlation-based Feature Selection (CFS) is used to select the best feature among 13 features extracted. The performance of a new feature set is called Feature Set 3 compared to CFS. The result shows that the proposed feature set achieves the highest level of accuracy.

TGS2611 performance as biogas monitoring instrument in digester model application

Gas chromatography for biogas concentration measurement is too expensive for people who made biogas independently. An instrument with cheaper TGS2611 sensor can be made to monitor biogas concentration. The instrument requires initialization for 60 seconds. TGS2611 sensor output voltage has proportionally changed to the temperature change. Two gas sensors used in the research have 12.125 mV/°C and 12.875 mV/°C of temperature constant. The instrument needs 10 seconds to respond the increase of biogas concentration and needs 25 seconds to respond the decrease of biogas concentration. The instrument which applied for biogas formation monitoring shows that higher temperature on digester B with average temperature 34.48°C, produces more biogas compared to digester A with average temperature 26.98°C after 34 hours of monitoring. Statistically, the instrument can be said has good precision due to its relatively small deviation, in average, gas sensor A has 1.36% and gas sensor B has 1.23% of relative standard deviation.

Classification of the syllables sound using wavelet, Renyi entropy and AR-PSD features

Feature extraction plays a very important role in the speech classification process because a better feature is good for improving the classification rate. This paper presents a speech feature extraction method by using Discrete Wavelet Transform (DWT) at 7th level of decomposition with mother wavelet of Dau-bechies 2, Renyi Entropy (RE), Autoregressive Power Spectral Density (AR-PSD), Statistical, as well as the combination of each method for extracting and classifying the certain Indonesian velar-vowel and alveolar-vowel syllables. Five different features set used in this study, namely the combination features of DWT and statistical (WS), RE, the combination of AR-PSD and Statistical (PSDS), the combination of PSDS and the selected features of RE (RPSDS), and the combination of DWT, RE, and AR-PSD (WRPSDS). Each syllable is segmented at a certain length to form a consonant-vowel. Multi-layer perceptron is used as a classifier after feature extraction process. The results show that the rank of the average recognition rate are WRPSDS, WS, RPSDS, PSDS, and RE, respectively.

Effective method for circle detection based on transformation of chain code direction

There are two approaches generally used for detecting a circle object in a digital image, they are a circle detection method based on Hough Transform and a circle detection method based on geometrical characteristics. A circle detection method on the basis of object geometry based on edges tracking of objects and transformation of chain code direction is proposed. Firstly, the chain code of the edges of the object to be detected was traced and stored in a set of vector array. Then the process of selecting the edges of objects from a bunch of the chain code was conducted to select circle candidates by focusing on changing the chain code direction on each edge of the object. Finally, the process of determining whether the object edges are circle candidate or not was carried out. During the process, the determination of reference point calculation was very influential. To increase efficiency at that stage was conducted by selecting three reference points on the edges of the object based on the calculation of...