• 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.

• Journal of Korea Multimedia Society
• /
• v.21 no.9
• /
• pp.1052-1061
• /
• 2018

In this paper, we propose a new algorithm that is robust against the effects of objects that are relatively unaffected by camera motion and can accurately detect camera motion even in high resolution images. First, for more accurate camera motion detection, a global motion filter based on entropy of a motion vector is used to distinguish the background and the object. A block matching algorithm is used to find exact motion vectors. In addition, a matched filter with the angle of the ideal motion vector of each block is used. Motion vectors including 4 kinds of diagonal direction, zoom in, and zoom out are added additionally. The experiment shows that the precision, recall, and accuracy of camera motion detection compared to the recent results is improved by 12.5%, 8.6% and 9.5%, respectively.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1996.06a
• /
• pp.195-198
• /
• 1996

In this paper, we introduce the notion of the smoothly perfect 8-connected (SP8C) contour and propose a coding technique for the SP8C contours. Based on the contour simplification using the majority filter proposed by Gu[6], SP8C contours are extracted on the contour lattice from the segmented image. By noting that, unlike the perfect 8-connected contours, the SP8C contours are restricted to travel in only 3 different directions along the contours, we also propose two techniques for encoding the SP8C contours. The one is the modified version of the neighbouring direction segment coding by Kandeko[2], while the other is to employ the notion of the entropy coding. From the comparison in terms of the entropy, it is shown that the proposed SP8C contours require les bits in encoding the diagonal contours than the 4-connected contours employed by Gu. And computer simulations reveal that the contours can be efficiently encoded by the proposed technique.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.54 no.1
• /
• pp.88-95
• /
• 2017

In this paper, we present a saliency detection method using entropy weight and Weber contrast in the wavelet transform domain. Our method is based on the commonly exploited conventional algorithms that are composed of the local bottom-up approach and global top-down approach. First, we perform the multi-level wavelet transform for the CIE Lab color images, and obtain global saliency by adding the local Weber contrasts to the corresponding low-frequency wavelet coefficients. Next, the local saliency is obtained by applying Gaussian filter that is weighted by entropy of wavelet high-frequency subband. The final saliency map is detected by non-lineally combining the local and global saliencies. To evaluate the proposed saliency detection method, we perform computer simulations for two image databases. Simulations results show the proposed method represents superior performance to the conventional algorithms.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.7
• /
• pp.157-168
• /
• 2014

We propose an novel artificial visual attention model that is capable of automatic detection and segmentation of saliency region on natural images in this paper. The proposed model is based on human visual perceptions in biological vision and contains there are main contributions. Firstly, we propose a novel framework of artificial visual attention model based on the opponent process theory using intensity and color features, and an entropy filter is designed to perceive salient regions considering the amount of information from intensity and color feature channels. The entropy filter is able to detect and segment salient regions in high segmentation accuracy and precision. Lastly, we also propose an adaptive combination method to generate a final saliency map. This method estimates scores about intensity and color conspicuous maps from each perception model and combines the conspicuous maps with weight derived from scores. In evaluation of saliency map by ROC analysis, the AUC of proposed model as 0.9256 approximately improved 15% whereas the AUC of previous state-of-the-art models as 0.7824. And in evaluation of salient region segmentation, the F-beta of proposed model as 0.7325 approximately improved 22% whereas the F-beta of previous state-of-the-art models.

• Journal of Korea Multimedia Society
• /
• v.5 no.2
• /
• pp.147-157
• /
• 2002

Noise in image sequences is visually offensive and may mask important image detail. In addition to degradation of visual quality, the noise pattern increases the entropy of the image, and thus hinders effective compression. This paper proposes a spatial and a temporal joint filters to reduce the noise by jointly connecting two adaptive noise reducers with different characteristics, and we also propose an IIR-type 3D noise reduction litter scheme connecting the spatial and the temporal joint filters. The proposed 3D IIR filter not only strongly removes noise in uniform image regions while preserving edges and details but also effectively suppresses temporal flicker caused by noise. Experimental results show that the proposed scheme improves subjective quality as well as objective quality as compared with the various noise filtering techniques.

• The Journal of Korea Robotics Society
• /
• v.2 no.1
• /
• pp.29-39
• /
• 2007

We present an implementation of particle filter algorithm for global localization and kidnap recovery of mobile robot. Firstly, we propose an algorithm for efficient particle initialization using sonar line features. And then, the average likelihood and entropy of normalized weights are used as a quality measure of pose estimation. Finally, we propose an active kidnap recovery by adding new particle set. New and independent particle set can be initialized by monitoring two quality measures. Added particle set can re-estimate the pose of kidnapped robot. Experimental results demonstrate the capability of our global localization and kidnap recovery algorithm.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.5 no.4
• /
• pp.237-242
• /
• 2005

H.264/AVC adopts new features compared with previous multimedia algorithms. It is inefficient to implement some of the new blocks using existing DSP instructions. Hence, new instructions are required to implement H.264/AVC. This paper proposes novel instructions for intra-prediction, in-loop deblocking filter, entropy coding and integer transform. Performance comparisons show that the required computation cycles for the in-loop deblocking filter can be reduced about $20{\sim}25%$. This paper also proposes new instructions for the integer transform. The proposed instructions can execute one dimension forward/inverse integer transform. The integer transform can be implemented using much smaller hardware size than existing DSPs.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.08a
• /
• pp.253-256
• /
• 2000

In this paper, we propose the lossless image compression algorithm using the integer wavelet transform. Recently, the S＋P transform is widely used and computed with only integer addition and bit-shift operations, but not proper to remove the correlation of smooth images. then we compare the Harr wavelet of the S＋P transform with various integer coefficient filter banks and apply 9/7 ICFB to the wavelet transform. In addition, we propose a entropy-coding method that exploits the multiresolution structure and the feedback of the prediction error, and can efficiently compress the transformed image for progressive transmission. Simulation results are compared to the compression ratio using the S+P transform with different types of images.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.4B
• /
• pp.715-724
• /
• 2000

This Paper presents 3-D subband filter banks which are effective for progressive and lossless compression of volumetric images. For such a purpose, ORT(Overlapping Rounding Transform), applied so far to 1-D losslesssubband filter banks, is now used to implement two types of 3-D lossless subband filter banks: separable andnon-separable types. Separable fiter banks are implemented form applying 1-D lossless filter banks consecutively.Non-separable later banks are developed by expanding the 1-D ORT into 3-D one. In particular, the proposed ORT based 3-D non-separable filter banks generalizes the 3-D HINT(Hierarchical INTerpolation) algorithm.Through the experiment comparisons on various volumetric medical images, we prove that the proposedseparablefnon-separable filter banks perform better, in terms of compression ratio (first order entropy), than theother lossless compression techniques such as block based transform and conventional 3-D HINT.