Yuita Arum Sari

Eye Movement as Navigator for Disabled Person

Eyes is one of human organs which mostly still functions properly in disabled people when other parts of the body are disabled. This research propose a new framework to recognize and detected eye movement for handling position by considering the decision of both left and right eye. The sophisticated algorithm, Haar Cascade Algorithm was used for observing the area of eyes, then thresholding image using morphology is used to obtain the focus of eyes. The Hough Circle Transform with several rules could decide the handling position of eye movement. The performance of the pro-posed algorithm could reach over 80% in all dataset.

A laser-vision based obstacle detection and distance estimation for smart wheelchair navigation

The aim of the research is to present an approach of obstacle distance estimation and navigation for smart wheelchair. The smart wheelchair is an electric wheelchair equipped with camera and line laser to navigate and avoid an obstacle. The camera was used to capture images from the environment to sense the pathway condition. The line laser was used in combination with camera to recognize an obstacle in the pathway based on the shape of line laser image in certain angle. A blob method detection was applied on the line laser image to recognize the pattern of the detected obstacles. The line laser projector and camera were mounted in fixed-certain position to make sure a fixed relation between blobs-gaps and obstacle-to-wheelchair distance. A simple linear regression from 16 obtained data was used to represent this relation as the estimated obstacle distance. As a result, the average error between the estimation and actual distance was 1.25 cm from 7 data testing experiments. The experiments result indicates that the proposed method is able to estimate well the distance between wheelchair and obstacle. Later, the smart wheelchair needs to decide further action whether it should turn left, right or just walk straight when facing certain obstacle to avoid it.

Color correction using improved linear regression algorithm

Color correction is one of essential stages in image processing, which plays an important role during image acquisition or pre-processing to produce a better color quality, before being used in further process. This paper proposes a new method for color correction using an improved linear regression algorithm based on a stepwise model. This proposed method is designed for assessing a series of discrete color levels, for instance in a leaf color chart. Color chart as a reference image is used for controlling color levels of a captured image or calibrating the image sensor. The experiment is conducted in L*a*b* color space, therefore a transformation from RGB into L*a*b* is needed at the first phase. The best matched color level between reference and captured image will be selected by k-Means clustering method. Chosen color levels are used for constructing linear regression function. This function is applied as well for removing outlier among color levels. To ensure the result of this color correction does not depend on lighting condition, the color constancy algorithm is acquired. Gray World and White Patch are chosen for color constancy methods. Compared to ordinary linear regression and color correction without adding color constancy, the combination of Gray World and improved linear regression algorithm based on stepwise model shows the best result in almost entire datasets in various lighting conditions.

Enhancing tomato clustering evaluation using color correction with improved linear regression in preprocessing phase

Color inconsistency poses many difficulties when capturing the same object using different image capture devices. Color is one of main parts in image preprocessing and therefore color correction is needed to calibrate images in order to produce consistent color values. In this paper, we propose a new color correction method by employing combined linear regression with stepwise model to enhance the quality of tomatoes ripeness clustering. Macbeth ColorChecker is needed as a reference image while a test image to be corrected is captured by an Android smartphone camera. There are 12 color levels to be compared between reference and test image. However, only a number of color levels are selected by k-means clustering. The selected color levels are utilized to build a linear regression algorithm with stepwise model. The result confirms that color correction and color constancy increase the clustering performance by 10% up to 40% for all possible configurations.

Tomato ripeness clustering using 6-means algorithm based on v-channel otsu segmentation

Segmentation process in an essential part in image processing to obtain good preparation either for further process of data mining or object recognition. This paper proposes a new method of segmenting tomato image for clustering its ripeness. The tomato images are taken from three types of smartphone camera in various lighting condition with white background. When taking picture by using smartphone camera, the image is a bit darker or lighter in certain side, so the segmentation is involved to the following stage. Color transformation is needed at the first stage of preprocessing which converts RGB channel to YUV channel in order to apply histogram equalization. YUV is better to perceptual similarities in machine vision than RGB. Histogram equalization is applied in single Y channel of an image. Afterwards merge a V channel to YUV channel then transform it to RGB color model to observe the difference and convert it back to YUV for segmentation. Otsu combined with V channel thresholding is utilized to segment image better. To evaluate the segmentation performance, clustering method is computed based on retrieved color of segmented image using K-Means, in which k=6 because of there are 6 stages of tomato ripeness. Color feature extraction by means of R, G, a∗, and b∗ color channel are treated subsequently. Experimental results show the system yield 1% Mean Square Error in clustering the ripeness of tomatoes.

Development of computer vision based obstacle detection and human tracking on smart wheelchair for disabled patient

People with physical disability such as quadriplegics may need a device which assist their mobility. Smart wheelchair is developed based on conventional wheelchair and is also generally equipped with sensors, cameras and computer based system as main processing unit to be able to perform specific algorithm for the intelligent capabilities. We develop smart wheelchair system that facilitates obstacle detection and human tracking based on computer vision. The experiment result of obstacle distance estimation using RANSAC showed lower average error, which is only 1.076 cm compared to linear regression which is 2.508 cm. The average accuracy of human guide detecting algorithm also showed acceptable result, which yield over 80% of accuracy.

Optimizing K-means text document clustering using latent semantic indexing and pillar algorithm

Document clustering is an important tool to help managing the vast amount of digital text document. This paper introduces a new approach to cluster text document. First, text is preprocessed and indexed using inverted index. Then the index is trimmed using TF-DF thresholding. After that, Term Document Matrix is built based on TF-IDF. Next step uses Latent Semantic Indexing to extract important feature from Term Document Matrix. The following process is selecting seeds via Pillar algorithm. Based on determined seeds, K-Means clustering is performed. Experiment result proves that this approach outperforms standard K-Means document clustering.