Sang-Hoon Kim

Optimal keyframe selection algorithm for three-dimensional reconstruction in uncalibrated multiple images

A keyframe selection algorithm is the process of selecting the essential image for 3-D reconstruction from among many uncalibrated images. Camera autocalibration between images is also essential for 3-D reconstruction. This paper proposes a keyframe selection algorithm that selects the best image to reduce reprojection error. The camera projection matrix in the selected keyframe arises from a full camera autocalibration process. From the camera projection matrix, which is assumed exact, the algorithm calculates the fundamental matrix using algebraic derivation. By this process, false matching is eliminated and finally a 3-D data set is obtained. In our experimental results, the proposed algorithm needs less time than other algorithms; it also has fewer errors in the reconstructed 3-D data. The fundamental matrix that is gained from algebraic derivation takes less time than other algorithms, and it has same average error as the others.

An Improved ICP Algorithm Based on the Sensor Projection for Automatic 3D Registration

Three-dimensional (3D) registration is the process aligning the range data sets form different views in a common coordinate system. In order to generate a complete 3D model, we need to refine the data sets after coarse registration. One of the most popular refinery techniques is the iterative closest point (ICP) algorithm, which starts with pre-estimated overlapping regions. This paper presents an improved ICP algorithm that can automatically register multiple 3D data sets from unknown viewpoints. The sensor projection that represents the mapping of the 3D data into its associated range image and a cross projection are used to determine the overlapping region of two range data sets. By combining ICP algorithm with the sensor projection, we can make an automatic registration of multiple 3D sets without pre-procedures that are prone to errors and any mechanical positioning device or manual assistance. The experimental results demonstrated that the proposed method can achieve more precise 3D registration of a couple of 3D data sets than previous methods.

Eye Localization based on Multi-Scale Gabor Feature Vector Model

Eye localization is necessary for face recognition and related application areas. Most of eye localization algorithms reported thus far still need to be improved about precision and computational time for successful applications. In this paper, we propose an improved eye localization method based on multi-scale Gator feature vector models. The proposed method first tries to locate eyes in the downscaled face image by utilizing Gabor Jet similarity between Gabor feature vector at an initial eye coordinates and the eye model bunch of the corresponding scale. The proposed method finally locates eyes in the original input face image after it processes in the same way recursively in each scaled face image by using the eye coordinates localized in the downscaled image as initial eye coordinates. Experiments verify that our proposed method improves the precision rate without causing much computational overhead compared with other eye localization methods reported in the previous researches.

Three-dimensional reconstruction of an object using three-planar homography of a single image

Camera calibration is essential in order to perform 3-D reconstruction from uncalibrated images. This paper proposes a new technique of camera calibration using a homography among three planar patterns taken from a single image. It is implemented more easily and simply than conventional methods. Experimental results support the hypothesis and show that the performance of the proposed method is better than that of conventional methods. Examples of 3-D reconstruction are demonstrated using the proposed algorithm on sequences of images.

Full three-dimensional reconstruction using key-frame selection under circular motion

Fundamental-matrix and key-frame selection constitute one of the most important techniques for full 3-D reconstruction of objects from turntable sequences. This paper proposes the new methods for these selection problems in 3-D reconstruction from uncalibrated sequences taken with a turntable and the Fotonovo camera system. Also, we propose a projection-matrix refinement for accurate full 3-D reconstruction. Our approach utilizes single-axis motion. To evaluate the fundamental matrix, camera calibration and 3-D registration are generally needed. Our main contribution is a method for robustly determining the corresponding points between two images, and for accurately filling gaps in a sparse object so as to make surface reconstruction tractable. We do not need all frames, but only few pairs of images (key frames). The key-frame selection has the advantage in camera pose estimation and 3-D scene reconstruction of reducing the computational costs. Experimental results on real image sequences demonstrate accurate object reconstructions and robustness of the proposed methods.

An Inductive-coupling Link with a Complementary Switching Transmitter and an Integrating Receiver

A transceiver for a high-speed inductivecoupling link is proposed. The bi-phase modulation (BPM) signaling scheme is used due to its good noise immunity. The transmitter utilizes a complementary switching method to remove glitches in transmitted data. To increase the timing margin on the receiver side, an integrating receiver with a pre-charging equalizer is employed. The proposed transceiver was implemented via a 130-nm CMOS process. The measured timing window for a 10 -12 bit error rate (BER) at 1.8 Gb/s was 0.33 UI.

An Efficient Illumination Preprocessing Algorithm based on Anisotropic Smoothing for Face Recognition

Robust face recognition under various illumination environments is very difficult and needs to be accomplished for successful commercialization. In this paper, we propose an efficient illumination preprocessing method for face recognition. illumination preprocessing algorithm based on anisotropic smoothing is well known to be effective among illumination normalization methods but deteriorates the intensity contrast of the original image, and incurs less sharp edges. The proposed method in this paper improves the previous anisotropic smoothing based illumination normalization method so that it increases the intensity contrast and enhances the edges while diminishing effects of illumination. Due to the result of these improvements, face images preprocessed by the proposed illumination preprocessing method becomes to have more distinctive feature vectors(Gabor feature vectors). Through experiments of face recognition using Gabor jet similarity, the effectiveness of the proposed illumination preprocessing method is verified.

3D-based synthesis and 3D reconstruction from uncalibrated images

In this paper, we propose a new camera calibration method for 3D-based image synthesis and 3D reconstruction. We improve the problem by changing the principle point for obtaining the linear equation. According to the error rate, we adapt the non-linear method that minimizes the intrinsic parameters. Namely, it minimizes the intrinsic parameters error by maintaining the computational conciseness. As a result, we can find optimized camera intrinsic parameters and adapt them for image synthesis and reconstruction.

Forward-view transition algorithm between the near and the far views and its extension to the panoramic images

We propose a new algorithm that generates intermediate scenes from the near and the far views. Assuming that two cameras can be approximated with the affine model, we align the optical axes of the near and the far views on the transition baseline using the Fourier method. Then we determine the scale of the desired novel scene, zoom in the far view, and zoom out the near view. The zoom-in far view is theoretically the same as the zoom-out near view, but disparities exist that cause the synthesized image to present the double-blurring effect. Therefore, we remove the disparities of the zoom-out near view and then synthesize it with the zoom-in far view. For the panorama-based navigation system, the proposed algorithm can also generate images having smooth and realistic transitions as the viewer walks from the current panoramic image to the next panoramic image.

3D Reconstruction using vanishing points

This paper proposes a calibration method from two images. Camera calibration is necessarily required to obtain 3D Information from 2D images. Previous works to accomplish the camera calibration needed the calibration object or required more than three images to calculate the Kruppa equation, however, we use the geometric constraints of parallelism and orthogonality can be easily presented in man-made scenes. The task of it is to obtain intrinsic and extrinsic camera parameters. The intrinsic parameters are evaluated from vanishing points and then the extrinsic parameters which are consisted of rotation matrix and translation vector of the camera are estimated from corresponding points of two views. From the calibrated parameters, we can recover the projection matrices for each view point. These projection matrices are used to recover 3D information of the scene and can be used to visualize new viewpoints.