• Journal of Electrical Engineering and Technology
• /
• v.10 no.4
• /
• pp.1877-1885
• /
• 2015

This paper presents a new Nearest-Neighbors based weighted representation for images and weighted K-Nearest-Neighbors (WKNN) classifier to improve the precision of image classification using the Bag of Visual Words (BOVW) based models. Scale-invariant feature transform (SIFT) features are firstly extracted from images. Then, the K-means++ algorithm is adopted in place of the conventional K-means algorithm to generate a more effective visual dictionary. Furthermore, the histogram of visual words becomes more expressive by utilizing the proposed weighted vector quantization (WVQ). Finally, WKNN classifier is applied to enhance the properties of the classification task between images in which similar levels of background noise are present. Average precision and absolute change degree are calculated to assess the classification performance and the stability of K-means++ algorithm, respectively. Experimental results on three diverse datasets: Caltech-101, Caltech-256 and PASCAL VOC 2011 show that the proposed WVQ method and WKNN method further improve the performance of classification.

• Proceedings of the KIEE Conference
• /
• 2006.07d
• /
• pp.1931-1932
• /
• 2006

로봇이 어떤 물체를 인지하고 그 물체에 대해 어떤 작업을 하고자 할 때 특정 물체의 인식 문제, 3차원 정보를 획득하는 문제, 자세를 추정하는 문제 등 해결해야 될 문제들이 있다. 물체를 인식하는 과정에서는 주위 배경과 물체의 크기의 변화, 회전, 가려짐 등으로 인해 물체 인식을 어렵게 만드는 요소들이 있다. 2차원 이미지를 통해 3차원 정보를 추출하는 과정은 일반적으로 두 대의 카메라를 이용하여 스테레오 이미지를 통해 얻는다. 이 때 좌우 영상간의 매칭의 과정이 필요하다. 자세 추정의 문제는 카메라 좌표와 물체의 좌표간의 관계를 알아야 한다. Visual Servoing을 어렵게 만드는 많은 요인들이 있으며 본 논문에서는 물체의 크기, 회전, 이동에 불변인 디스크립터(descriptor)를 사용하는 SIFT(Scale Invariant Feature Transform)를 통해 3차원 물체의 인식과 자세를 추정하는 방법을 제시한다. 또한 자세 추정을 위해 2차원 Keypoint들의 매칭을 3차원 정보를 통해 검증하는 방법을 제시한다. (SIFT에 의해 추출된 point를 Keypoint라 명한다.)

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.13 no.2
• /
• pp.157-169
• /
• 2013

Scale Invariant Feature Transform (SIFT) generates image features widely used to match objects in different images. Previous work on hardware-based SIFT implementation requires excessive internal memory and hardware logic [1]. In this paper, a new hardware organization is proposed to implement SIFT with less memory and hardware cost than the previous work. To this end, a parallel Gaussian filter bank is adopted to eliminate the buffers that store intermediate results because parallel operations allow all intermediate results available at the same time. Furthermore, the processing order is changed from the raster-scan order to the block-by-block order so that the line buffer size storing the source image is also reduced. These techniques trade the reduction of memory size with a slight increase of the execution time and external memory bandwidth. As a result, the memory size is reduced by 94.4%. The proposed hardware for SIFT implementation includes the Descriptor generation block, which is omitted in the previous work [1]. The addition of the hardwired descriptor generation improves the computation speed by about 30 times when compared with the previous work.

• Journal of Broadcast Engineering
• /
• v.17 no.5
• /
• pp.838-850
• /
• 2012

In this paper, we propose a novel algorithm that generates multi-view images by using various image features obtained from the given stereoscopic images. In the proposed algorithm, we first create an intensity gradient saliency map from the given stereo images. And then we calculate a block-based optical flow that represents the relative movement(disparity) of each block with certain size between left and right images. And we also obtain the disparities of feature points that are extracted by SIFT(scale-invariant We then create a disparity saliency map by combining these extracted disparity features. Disparity saliency map is refined through the occlusion detection and removal of false disparities. Thirdly, we extract straight line segments in order to minimize the distortion of straight lines during the image warping. Finally, we generate multi-view images by grid mesh-based image warping algorithm. Extracted image features are used as constraints during grid mesh-based image warping. The experimental results show that the proposed algorithm performs better than the conventional DIBR algorithm in terms of visual quality.

• Journal of Biomedical Engineering Research
• /
• v.32 no.3
• /
• pp.237-244
• /
• 2011

In this paper, a feature-based registration technique is proposed for pre-contrast and post-contrast lung CT images. It utilizes three dimensional(3-D) features with their descriptors and estimates feature correspondences by nearest neighborhood matching in the feature space. We design a transformation model between the input image pairs using a free form deformation(FFD) which is based on B-splines. Registration is achieved by minimizing an energy function incorporating the smoothness of FFD and the correspondence information through a non-linear gradient conjugate method. To deal with outliers in feature matching, our energy model integrates a robust estimator which discards outliers effectively by iteratively reducing a radius of confidence in the minimization process. Performance evaluation was carried out in terms of accuracy and efficiency using seven pairs of lung CT images of clinical practice. For a quantitative assessment, a radiologist specialized in thorax manually placed landmarks on each CT image pair. In comparative evaluation to a conventional feature-based registration method, our algorithm showed improved performances in both accuracy and efficiency.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.43 no.3 s.309
• /
• pp.8-15
• /
• 2006

Obstacle detection is an important task for many mobile robot applications. The methods using stereo vision and optical flow are computationally expensive. Therefore, this paper presents a vision-based obstacle detection method using only two view images. The method uses a single passive camera and odometry, performs in real-time. The proposed method is an obstacle detection method using 3D reconstruction from taro views. Processing begins with feature extraction for each input image using Dr. Lowe's SIFT(Scale Invariant Feature Transform) and establish the correspondence of features across input images. Using extrinsic camera rotation and translation matrix which is provided by odometry, we could calculate the 3D position of these corresponding points by triangulation. The results of triangulation are partial 3D reconstruction for obstacles. The proposed method has been tested successfully on an indoor mobile robot and is able to detect obstacles at 75msec.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.3
• /
• pp.630-638
• /
• 2016

In this paper, a feature based panorama image generation algorithm using FAST(Features from Accelerated Segment Test) method that is faster than SIFT(Scale Invariant Feature Transform) and SURF(Speeded Up Robust Features) is proposed. Cylindrical projection is performed to generate natural panorama images with numerous images as input. The occurred error can be minimized by applying RANSAC(Random Sample Consensus) for the matching process. When we synthesize numerous images acquired from different camera angles, we use blending techniques to compensate the distortions by the heterogeneity of border line. In that way, we could get more natural synthesized panorama image. The proposed algorithm can generate natural panorama images regardless the order of input images and tilted images. In addition, the image matching can be faster than the conventional method. As a result of the experiments, distortion was corrected and natural panorama image was generated.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.9
• /
• pp.1972-1978
• /
• 2010

AAM(Active Appearance Model) is one of the most effective ways to detect deformable 2D objects and is a kind of mathematical optimization methods. The cost function is a convex function because it is a least-square function, but the search space is not convex space so it is not guaranteed that a local minimum is the optimal solution. That is, if the initial value does not depart from around the global minimum, it converges to a local minimum, so it is difficult to detect face contour correctly. In this study, an AAM-based face tracking algorithm is proposed, which is robust to various lighting conditions and backgrounds. Eye detection is performed using SIFT and Genetic algorithm, the information of eye are used for AAM's initial matching information. Through experiments, it is verified that the proposed AAM-based face tracking method is more robust with respect to pose and background of face than the conventional basic AAM-based face tracking method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.6
• /
• pp.2230-2250
• /
• 2015

Multiple projectors with partially overlapping regions can be used to project a seamless image on a large projection surface. With the advent of high-resolution photography, such systems are gaining popularity. Experts set up such projection systems by subjectively identifying the types of errors induced by the system in the projected images and rectifying them by optimizing (correcting) the parameters associated with the system. This requires substantial time and effort, thus making it difficult to set up such systems. Moreover, comparing the performance of different multi-projector display (MPD) systems becomes difficult because of the subjective nature of evaluation. In this work, we present a framework to quantitatively determine the quality of an MPD system and any image projected using such a system. We have divided the quality assessment into geometric and photometric qualities. For geometric quality assessment, we use Feature Similarity Index (FSIM) and distance-based Scale Invariant Feature Transform (SIFT). For photometric quality assessment, we propose to use a measure incorporating Spectral Angle Mapper (SAM), Intensity Magnitude Ratio (IMR) and Perceptual Color Difference (ΔE). We have tested the proposed framework and demonstrated that it provides an acceptable method for both quantitative evaluation of MPD systems and estimation of the perceptual quality of any image projected by them.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.38 no.5
• /
• pp.477-486
• /
• 2020

Drone has been widely used for many applications ranging from amateur and leisure to professionals to get fast and accurate 3-D information of the surface of the interest. Most of commercial softwares developed for this purpose are performing automatic matching based on SIFT (Scale Invariant Feature Transform) or SURF (Speeded-Up Robust Features) using nadir-looking stereo image sets. Since, there are some situations where not only nadir and nadir-looking matching, but also nadir and oblique-looking matching is needed, the existing software for the latter case could not get good results. In this study, a matching experiment was performed to utilize images with differences in geometry. Nadir and oblique-looking images were acquired through drone for a total of 2 times. SIFT, SURF, which are feature point-based, and IMAS (Image Matching by Affine Simulation) matching techniques based on affine transformation were applied. The experiment was classified according to the identity of the geometry, and the presence or absence of a building was considered. Images with the same geometry could be matched through three matching techniques. However, for image sets with different geometry, only the IMAS method was successful with and without building areas. It was found that when performing matching for use of images with different geometry, the affine transformation-based matching technique should be applied.