Yi Mou

A New Algorithm for Shoreline Extraction from Satellite Imagery with Non-Separable Wavelet and Level Set Method

 Abstract—An effective and precise method for shoreline detection from satellite imagery is presented. The algorithm is based on two main steps: (1)the detection of singularities in a single image using non-separable wavelet and (2)amendment procedure using distance regularized level set evolution scheme. Firstly, by selecting appropriate parameters, the non-separable wavelet filter banks which can provide information of different orientations are used to capture the singularities of the selected single satellite image; Secondly, obtaining the modulus image by utilizing sub-images decomposed from the non-separable wavelet filter banks; Thirdly, extracting the shoreline iteratively with the use of distance regularized level set evolution scheme. Experiments are conducted and results show that the proposed algorithm is applicable to satellite imagery, and the shoreline is robust to noises as well as blurring.

STFT-like time frequency representations of nonstationary signal with arbitrary sampling schemes

Spectrograms provide an effective way of time-frequency representation (TFR). Among these, short-time Fourier transform (STFT) based spectrograms are widely used for various applications. However, STFT spectrogram and its revised versions suffer from two main issues: (1) there is a trade-off between time resolution and frequency resolution and (2) almost all the existing TFR methods, including STFT spectrogram, are not designed to handle arbitrary nonuniformly sampled data. To address these two issues, short-time iterative adaptive approach (ST-IAA) was recently proposed as a data-dependent adaptive spectral estimation method that can provide much enhanced TFR performance. In this paper, inspired by the ST-IAA method, we present an alternative approach, namely short-time sparse learning via iterative minimization (ST-SLIM), which can provide sparser and slightly better TFR performance than its ST-IAA counterpart. Moreover, in order to extend the applicability of ST-IAA to signals in the missing data case, we also propose a short-time missing-data iterative adaptive approach (ST-MIAA) which can retrieve the missing data effectively and outperform ST-IAA and ST-SLIM in the missing data case. We will demonstrate via simulation results the superiority of our proposed algorithms in terms of resolution, sidelobe suppression and applicability to signals with arbitrary sampling patterns.

Single image rain streaks removal based on self-learning and structured sparse representation

Rain streaks removal from single image is a challenging problem for image processing. This paper proposed a novel algorithm for rain streaks removal to single image based on a self-learning framework and structured sparse representation. More precisely, our algorithm firstly segments and categorizes input image into “rain streaks” regions and “non-rain geometric” regions via texture analysis. Meanwhile, we also decompose input image into high-frequency (HF) and low-frequency (LF) parts with bilateral filtering. Followed that, we introduced our newly proposed structured dictionary learning to decompose HF part into “rain texture” details and “non-rain geometric” details, where patches for training rain and non-rain sub-dictionaries are automatically selected from “rain streaks” and “non-rain geometric” regions. Finally, we combine LF part with non-rain geometric details to get rain-streaks-removal image. Experiments demonstrate the superiority of our proposed algorithm.

Human Heart Rate Estimation Using Ordinary Cameras under Natural Movement

Non-contact face-video based human heart rate (HR) estimation has attracted a lot of attentions in recent years. Almost all the state-of-the-art webcam or smartphone based HR estimation methods comprise three main steps: firstly, a region of interest (ROI) on the human face is detected in each video frame, then, the target signal is obtained by fusing multiple raw traces, which are extracted from the RGB channels across all the video frames, finally, HR is estimated by applying frequency analysis approach to the target signal. However, three major drawbacks impede the applicability of the current methods: (1) the performance of ROI detection is susceptible to head motion and facial expression, (2) there is still a lack of well-accepted method for fusing raw traces to form the target signal, and (3) the adopted frequency analysis approaches always provide estimation results with low resolution and high side lobes. To address these issues, we propose a novel HR estimation method which is applicable to ordinary cameras subject to natural head movement or facial expression. The proposed method features ROI detection via facial feature detection and tracking, target signal extraction via Independent Component Analysis (ICA) in the RGB channels, and HR estimation via real-valued iterative adaptive approach (RIAA). Experimental results validate the superiority of our proposed method.

Webcam-Based Visual Gaze Estimation Under Desktop Environment

Image-based visual gaze estimation has been widely used in various scientific and application-oriented disciplines. However, the high cost and tedious calibration procedure impede its generalization in real scenarios. In this paper, we develop a low cost yet effective webcam based visual gaze estimation system. Different from previous works, we aim at minimizing the system cost, and at the same time, making the system more flexible and feasible to users. More specifically, only a single ordinary webcam is used in our system. Meanwhile, we also proposed a novel calibration mechanism which takes account binocular feature vectors simultaneously, and uses only four visual target points. We compare our system with the state of the art webcam based visual gaze estimation methods. Experimental results demonstrate that our system can achieve satisfactory performance without the requirements of dedicated hardware or tedious calibration procedure.

Locally Linear Embedding Based Dynamic Texture Synthesis

Dynamic textures are often modeled as a low-dimensional dynamic process. The process usually comprises an appearance model of dimension reduction, a Markovian dynamic model in latent space to synthesize consecutive new latent variables and a observation model to map new latent variables onto the observation space. Linear dynamic system(LDS) is effective in modeling simple dynamic scenes while is hard to capture the nonlinearities of video sequences, which often results in poor visual quality of the synthesized videos. In this paper,we propose a new framework for generating dynamic textures by using a new appearance model and a new observation model to preserves the non-linear correlation of video sequences. We use locally linear embedding(LLE) to create an manifold embedding of the input sequence, apply a Markovian dynamics to maintain the temporal coherence in the latent space and synthesize new manifold, and develop a novel neighbor embedding based method to reconstruct the new manifold into the image space to constitute new texture videos. Experiments show that our method is efficient in capturing complex appearance variation while maintaining the temporal coherence of the new synthesized texture videos.

Gaussian latent variable models for variable selection.

Variable selection has been extensively studied in linear regression and classification models. Most of these models assume that the input variables are noise free, the response variables are corrupted by Gaussian noise. In this paper, we discuss the variable selection problem assuming that both input variables and response variables are corrupted by Gaussian noise. We analyze the prediction error when augment one related noise variable. We show that the prediction error always decrease when more variable were employed for prediction when the joint distribution of variables are known. Based on this analysis, in sense of mean square error, the optimal variable selection can be obtained. We found that the results is very different from the matching pursuit algorithm(MP), which is widely used in variable selection problems.

STFT-like time frequency representations for nonstationary signal — From evenly sampled data to arbitrary nonuniformly sampled data

Spectrograms provide an effective way for time-frequency representation (TFR). Among these, short-time Fourier transform (STFT) based spectrograms are extensively used for various applications. However, STFT spectrogram and its revised versions suffer from two main issues: (1) there is a trade-off between time resolution and frequency resolution, and (2) almost all existing TFR methods, including STFT spectrogram, are not suitable to deal with nonuniformly sampled data. In this paper, we address these two problems by presenting alternative approaches, namely short-time amplitude and phase estimation (ST-APES) and short-time sparse learning via iterative minimization (ST-SLIM), to improve the resolution of STFT based spectrogram, and extend the applicability of our approaches to signals with arbitrary sampling patterns. Apart from evenly sampled data, we will consider missing data as well as arbitrary nonuniformly sampled data, at the same time. We will demonstrate via simulation results the superiority of our proposed algorithms in terms of resolution, sidelobe suppression and applicability to signals with arbitrary sampling patterns.

Global sparse partial least squares.

The partial least squares (PLS) is designed for prediction problems when the number of predictors is larger than the number of training samples. PLS is based on latent components that are linear combinations of all of the original predictors, it automatically employs all predictors regardless of their relevance. This will degrade its performance and make it difficult to interpret the result. In this paper, global sparse PLS (GSPLS) is proposed to allow common variable selection in each deflation process as well as dimension reduction. We introduce the ℓ2, 1 norm to direction matrix and develop an algorithm for GSPLS via employing the Bregmen Iteration algorithm, illustrate the performance of proposed method with an analysis to red wine dataset. Numerical studies demonstrate the superiority of proposed GSPLS compared with standard PLS and other existing methods for variable selection and prediction in most of the cases.

A New Approach for Spectra Baseline Correction using Sparse Representation

A new baseline correction algorithm for spectral signal based on sparse representation is proposed. Firstly, utilizing the training sample to obtain the dictionaries of both baseline and spectrum; Secondly, establishing sparse representation model of spectral signal; thirdly, employing OMP algorithm to calculate the representation coefficients of spectral signal and finally, obtaining the spectral baseline from representation coefficients which are corresponded to the baseline dictionary. Then, the spectra baseline correction is completed by removing the baseline from original observed spectrum. Contrast experiment and quantitative analysis of corrected spectral signals are conducted and results show the highly efficiency and accuracy of the proposed algorithm.