• The Journal of the Acoustical Society of Korea
• /
• v.34 no.4
• /
• pp.321-327
• /
• 2015

In this paper, we propose an algorithm for achieving robust Visual Voice Activity Detection (VVAD) for enhanced speech recognition. In conventional VVAD algorithms, the motion of lip region is found by applying an optical flow or Chaos inspired measures for detecting visual speech frames. The optical flow-based VVAD is difficult to be adopted to driving scenarios due to its computational complexity. While invariant to illumination changes, Chaos theory based VVAD method is sensitive to motion translations caused by driver's head movements. The proposed Local Variance Histogram (LVH) is robust to the pixel intensity changes from both illumination change and translation change. Hence, for improved performance in environmental changes, we adopt the novel threshold estimation using total variance change. In the experimental results, the proposed VVAD algorithm achieves robustness in various driving situations.

• Journal of KIISE:Databases
• /
• v.36 no.6
• /
• pp.446-454
• /
• 2009

Among various region detector and shape feature extraction method, MSER(Maximally Stable Extremal Region) and SIFT and its variant methods are popularly used in computer vision application. However, since SIFT is sensitive to the illumination change and MSER is sensitive to the scale change, it is not easy to apply the image similarity retrieval. In this paper, we present a Scale and Rotation Invariant Region Feature(SRIRF) descriptor using scale pyramid, MSER and affine normalization. The proposed SRIRF method is robust to scale, rotation, illumination change of image since it uses the affine normalization and the scale pyramid. We have tested the SRIRF method on various images. Experimental results demonstrate that the retrieval performance of the SRIRF method is about 20%, 38%, 11%, 24% better than those of traditional SIFT, PCA-SIFT, CE-SIFT and SURF, respectively.

• Journal of Information Processing Systems
• /
• v.11 no.4
• /
• pp.643-654
• /
• 2015

Face recognition under controlled settings, such as limited viewpoint and illumination change, can achieve good performance nowadays. However, real world application for face recognition is still challenging. In this paper, we propose using the combination of Affine Scale Invariant Feature Transform (SIFT) and Probabilistic Similarity for face recognition under a large viewpoint change. Affine SIFT is an extension of SIFT algorithm to detect affine invariant local descriptors. Affine SIFT generates a series of different viewpoints using affine transformation. In this way, it allows for a viewpoint difference between the gallery face and probe face. However, the human face is not planar as it contains significant 3D depth. Affine SIFT does not work well for significant change in pose. To complement this, we combined it with probabilistic similarity, which gets the log likelihood between the probe and gallery face based on sum of squared difference (SSD) distribution in an offline learning process. Our experiment results show that our framework achieves impressive better recognition accuracy than other algorithms compared on the FERET database.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2014.04a
• /
• pp.781-784
• /
• 2014

Face recognition under controlled settings, such as limited viewpoint and illumination change, can achieve good performance nowadays. However, real world application for face recognition is still challenging. In this paper, we use Affine SIFT to detect affine invariant local descriptors for face recognition under large viewpoint change. Affine SIFT is an extension of SIFT algorithm. SIFT algorithm is scale and rotation invariant, which is powerful for small viewpoint changes in face recognition, but it fails when large viewpoint change exists. In our scheme, Affine SIFT is used for both gallery face and probe face, which generates a series of different viewpoints using affine transformation. Therefore, Affine SIFT allows viewpoint difference between gallery face and probe face. Experiment results show our framework achieves better recognition accuracy than SIFT algorithm on FERET database.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.7 no.3
• /
• pp.196-208
• /
• 2007

An analytical model is proposed for an optically controlled Metal Semiconductor Field Effect Transistor (MESFET), known as Optical Field Effect Transistor (OPFET) considering the diffusion fabrication process. The electrical parameters such as threshold voltage, drain-source current, gate capacitances and switching response have been determined for the dark and various illuminated conditions. The Photovoltaic effect due to photogenerated carriers under illumination is shown to modulate the channel cross-section, which in turn significantly changes the threshold voltage, drainsource current, the gate capacitances and the device switching speed. The threshold voltage $V_T$ is reduced under optical illumination condition, which leads the device to change the device property from enhancement mode to depletion mode depending on photon impurity flux density. The resulting I-V characteristics show that the drain-source current IDS for different gate-source voltage $V_{gs}$ is significantly increased with optical illumination for photon flux densities of ${\Phi}=10^{15}\;and\;10^{17}/cm^2s$ compared to the dark condition. Further more, the drain-source current as a function of drain-source voltage $V_{DS}$ is evaluated to find the I-V characteristics for various pinch-off voltages $V_P$ for optimization of impurity flux density $Q_{Diff}$ by diffusion process. The resulting I-V characteristics also show that the diffusion process introduces less process-induced damage compared to ion implantation, which suffers from current reduction due to a large number of defects introduced by the ion implantation process. Further the results show significant increase in gate-source capacitance $C_{gs}$ and gate-drain capacitance $C_{gd}$ for optical illuminations, where the photo-induced voltage has a significant role on gate capacitances. The switching time ${\tau}$ of the OPFET device is computed for dark and illumination conditions. The switching time ${\tau}$ is greatly reduced by optical illumination and is also a function of device active layer thickness and corresponding impurity flux density $Q_{Diff}$. Thus it is shown that the diffusion process shows great potential for improvement of optoelectronic devices in quantum efficiency and other performance areas.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2004-2007
• /
• 2003

This paper describes a method of performance enhancement using Flatness Mesure(FM) for the Gaussian Mixture Model(GMM) face recognition systems. Using this measure we discard the frames having low information before training and test. As the result, the performance increases about 9％ in the lower mixtures and calculation burden is decreased. As well, the recognition error rate is decreased under the illumination change surroundings. We use the 2D DCT coefficients lot face feature vectors and experiments are carried out on the Olivetti Research Laboratory (ORL) face database.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.6
• /
• pp.1215-1224
• /
• 2011

Recently, a lot of research for the smart grid technology have been carried out to achieve energy efficiency for the electronic products. In order to practically apply this study, smart instruments which are capable of the AMI (Advanced Metering Infrastructure) and DR (Demand Response) function are necessary. However, it is difficult to apply the function of the smart grid to the electronic product that cannot support the smart grid. Accordingly, the efficient use of electric energy is impossible. In order to solve this problem, the electronic product has to be changed into the exclusive electronic product supporting smart grid technology or the smart controller has to be attached the outside of the device. In this study, we developed the smart controller for connecting the electric appliances to the smart grid system. It can be attached to the illumination and the smart grid-based lamp control system at home. We additionally designed the message frame and the protocol to operate the smart controller with the AMI based EMS (Energy Management Server). We developed an experimental system to practically verify functions of the smart controller which is attached to the lighting device. From the system, we showed that the electric source of the illumination can be controlled according to the load change and saved energy effectively. We also confirmed the structural benefit and the energy-efficient effect through the verification of the smart controller.

• Journal of Broadcast Engineering
• /
• v.18 no.1
• /
• pp.77-87
• /
• 2013

In this paper, we present a novel method for local image descriptor called exact order based descriptor (EOD) which is robust to illumination changes and Gaussian noise. Exact orders of image patch is induced by changing discrete intensity value into k-dimensional continuous vector to resolve the ambiguity of ordering for same intensity pixel value. EOD is generated from overall distribution of exact orders in the patch. The proposed local descriptor is compared with several state-of-the-art descriptors over a number of images. Experimental results show that the proposed method outperforms many state-of-the-art descriptors in the presence of illumination changes, blur and viewpoint change. Also, the proposed method can be used for many computer vision applications such as face recognition, texture recognition and image analysis.

• Journal of Advanced Marine Engineering and Technology
• /
• v.38 no.9
• /
• pp.1112-1118
• /
• 2014

In this paper, we propose a local descriptor and a similarity measure that is robust to radiometic variations. The proposed local descriptor is made up Haar wavelet filter and it can contain frequency informations about the feature point and its surrounding pixels in fixed region, and it is able to describe feature point clearly under ununiform illumination condition. And a proposed similarity measure is combined with conventional entropy-based similarity and another similarities that is generated by local descriptor. It can reflect similarities between image regions accurately under radiometic illumination variations. We validate with experimental results on some images and we confirm that the proposed algorithm is more superior than conventional algorithms.

• The KIPS Transactions:PartD
• /
• v.13D no.2 s.105
• /
• pp.287-292
• /
• 2006

There are two problems associated with the existing principal component analysis, which is regarded as the most effective in object recognition technology. First, it brings about an increase in the volume of calculations in proportion to the square of image size. Second, it gives rise to a decrease in accuracy according to illumination changes. In order to solve these problems, this paper proposes wavelet transformation and histogram equalization. Wavelet transformation solves the first problem by using the images of low resolution. To solve the second problem the histogram equalization enlarges the contrast of images and widens the distribution of brightness values. The proposed technology improves recognition rate by minimizing the effect of illumination change. It also speeds up the processing and reduces its area by wavelet transformation.