Dong-Min Woo

Error detection and DEM fusion using self-consistency

The ability to efficiently and robustly recover accurate 3D terrain models from sets of stereoscopic images is important to many civilian and military applications. Our long-term goal is to develop an automatic, multi-image 3D reconstruction algorithm that can be applied to these domains. To develop an effective and practical terrain modeling system, methods must be found for detecting unreliable elevations in digital elevation maps (DEMs), and for fusing several DEMs from multiple sources into an accurate and reliable result. This paper focuses on two key factors for generating robust 3D terrain models, (1) the ability to detect unreliable elevations estimates, and (2) to fuse the reliable elevations into a single optimal terrain model. The techniques discussed in this paper are based on the concept of using self-consistency to identify potentially unreliable points. We apply the self-consistency methodology to both the two-image and multi-image scenarios. We demonstrate that the recently developed concept of self-consistency can be effectively employed to determine the reliability of values in a DEM. Estimates with a reliability below an error threshold can be excluded from further processing. We test the effectiveness of the methodology, as well as the relationship between error rate and scene geometry by processing both real and photo-realistic simulations.

Medical Image Authentication Using Hash Function and Integer Wavelet Transform

In this paper, a new fragile watermarking algorithm for medical images is proposed. This algorithm makes it possible to resolve the security and forgery problem of the medical images. Instead of the discrete wavelet transform, an integer wavelet transform is used to utilize hash function. The watermark associated with the hash values is inserted into the LSBs of the integer wavelet transform coefficients. From the experimental results, it can be confirmed that the proposed algorithm detects a forged area of the image very well.

A blind watermarking algorithm based on HVS and RBF neural network for digital image

This paper proposes a new blind watermarking scheme in which a watermark is embedded into the discrete wavelet transform (DWT) domain. The method uses the HVS model, and radial basis function neural networks (RBF). RBF will be implemented while embedding and extracting watermark. The human visual system (HVS) model is used to determine the watermark insertion strength. The inserted watermark is a random sequence. The secret key determines the beginning position of the image where the watermark is embedded. This process prevents possible pirates from removing the watermark easily. Experimental results show that the proposed method has good imperceptibility and high robustness to common image processing attacks.

3D Building Detection and Reconstruction from Aerial Images Using Perceptual Organization and Fast Graph Search

This paper presents a new method for building detection and reconstruction from aerial images. In our approach, we extract useful building location information from the generated disparity map to segment the interested objects and consequently reduce unnecessary line segments extracted in the low level feature extraction step. Hypothesis selection is carried out by using an undirected graph, in which close cycles represent complete rooftops hypotheses. We test the proposed method with the synthetic images generated from Avenches dataset of Ascona aerial images. The experiment result shows that the extracted 3D line segments of the reconstructed buildings have an average error of 1.69m and our method can be efficiently used for the task of building detection and reconstruction from aerial images.

An Efficient Method for Camera Calibration Using MultiLayer Perceptron Type Neural Network

This paper presents a 3D camera calibration method based on a nonlinear modeling function of an artificial neural network. The neural network employed in this paper is primarily used as a nonlinear mapper between 2D image points and points of a certain space in 3D real world. The neural network model implicitly contains all the physical parameters, some of which are very difficult to be estimated in the conventional calibration methods. MutiLayer Perceptron Type Neural Network (MLPNN) is employed to implement the relationship between image coordinates. In order to show the performance of the proposed method, we carry out experiments on the estimation of 2D image coordinates given 3D real world coordinates. The experimental results show that the proposed method improved calibration accuracy over widely used Tsais two stage method (TSM).

Centroid neural network with Chi square distance measure for texture classification

An unsupervised competitive neural network for efficient classification of image textures is proposed. The proposed neural network architecture, called centroid neural network with Chi square distance measure (CNN-χ2), employs the Chi square measure as its distance measure and utilizes the local binary pattern (LBP) as an effective feature extraction tool for image data. The proposed CNN-χ2 is applied to image texture classification problems on the Brodatz texture album database. The results are compared with those of conventional approaches including the HMT (hidden Markov tree), IMM (independence mixture model), and WES (wavelet energy signatures). The evaluated results demonstrate that the proposed CNN-χ2 classification algorithm outperforms the conventional algorithms in terms of classification accuracy.

Image classification using Gradient-Based Fuzzy c-Means with Divergence Measure

This paper proposes a novel classification method for image retrieval using gradient-based fuzzy c-means with divergence measure (GBFCM(DM)). GBFCM(DM) is a neural network-based algorithm that utilizes the Divergence Measure to exploit the statistical nature of the image data and thereby improve the classification accuracy. Experiments and results on various data sets demonstrate that the proposed classification algorithm outperforms conventional algorithms such as the traditional self-organizing map (SOM) and fuzzy c-means (FCM) by 27% - 28.5% in terms of accuracy.

Robust digital image watermarking algorithm using BPN neural networks

This paper proposes a new watermarking scheme in which a logo watermark is embedded into the spatial domain of the original image using Back-Propagation neural networks (BPN). BPN will learn the characteristic of the image, and then watermark is embedded and extracted by the trained BPN. The image is divided into 8(8 blocks and the average pixel value of each block is used as the desired output value of the BPN. The quantized DC coefficient of discrete cosine transform (DCT) domain of each block is used as input value of the BPN to be trained. After the BPN is trained using those input/output values, watermark is embedded into the spatial domain using the trained BPN. The trained BPN also used in watermark extracting process. Experimental results show that the proposed method has good imperceptibility and high robustness to common image processing.

Prediction of Sunspot Series Using BiLinear Recurrent Neural Network

A prediction scheme of sunspot series using a BiLinear Recurrent Neural Network (BLRNN) is proposed in this paper. Since the BLRNN is based on the bilinear polynomial, it has been successfully used in modeling highly nonlinear systems with time-series characteristics and the BLRNN can be a natural choice in predicting sunspot series. The performance of the proposed BLRNN-based predictor is evaluated and compared with the conventional MultiLayer Perceptron Type Neural Network (MLPNN)-based predictor. Experiments are conducted on the Wolf sunspot series number data. The results show that the proposed BLRNN based predictor outperforms the MLPNN-based one interms of the Normalized Mean Squared Error (NMSE).

Extraction of 3D Line Segment Using Digital Elevation Data

3D line segment can be regarded as one of the most useful features in constructing 3D model. In this context, this paper presents a new 3D line segment extraction method by using line fitting of elevation data on 2D line coordinates of ortho-image. In order to use elevation in line fitting, the elevation itself should be reliable. To measure the reliability of elevation, in this paper, we employ the concept of self-consistency. We test the effectiveness of the proposed method with a quantitative accuracy analysis using synthetic images generated from Avenches data set of Ascona aerial images. Experimental results indicate that our method generates 3D line segments almost 10 times more accurate than raw elevations obtained by area-based method.