ROI를 이용한 웨이브렛 기반 디지털 영상의 워터마킹 기법

본 논문에서 디지털 워터마킹을 위해 제안한 방법은 다중 해상도 웨이블릿 변환을 기본으로 하고 있다. 영상 데이터의 저작권 보호를 위해 필요한 워터마크로 <TEX>$2N_{wx}{\times}2N_{wy}$</TEX> 크기의 이진 로고영상의 1 레벨 DWT(discrete wavelet transform) 계수값을 사용하였다. 대상 영상을 3 레벨 DWT한 후 LL 영역과 중주파수 대역을 <TEX>$N_{wx}{\times}N_{wy}$</TEX> 크기로 분할하고, 분할된 블록내 값이 큰 계수값들을 이용하여 임계값을 설정한다. 대상 영상의 각 주파수 대역마다 설정되는 임계값은 동일 대역의 워터마크가 다 삽입될 수 있는 값을 기준으로 정한다 즉, 각 주파수 대역마다 임계값을 설정해야 한다. 대상 영상의 화질 저하를 막기 위해 블록내 임계값 이상인 부분에 대해 워터마크의 동일 위치 값을 삽입한다. 워터마크를 대상 영상의 LL 영역 및 중주파수 대역에 삽입한다. 워터마크를 인간의 시각 시스템으로부터 감추기 위해 워터마크에 HVS(human visual system)을 적용하여 삽입하였다 따라서 본 논문에서 제안한 기법은 워터마크의 빠르고 정확한 검출이 가능하며, HVS가 적용된 작은 크기의 워터마크를 삽입함으로써 비가시성과 강건성이 뛰어나다는 장점이 있다. In this paper. the proposed method for the digital watermarking is based on the multiresolution wavelet transform. The 1-level Discrete Wavelet Transform(DWT) coefficients of a <TEX>$2N_{wx}{\times}2N_{wy}$</TEX> binary logo image used as a watermarks. The LL band and middle frequency band of the host image that the 3-level DWT has been performed are divided into <TEX>$N_{wx}{\times}N_{wy}$</TEX> size and we use large coefficients at the divided blocks to make threshold. we set the thresholds that completely insert the watermark in each frequency of the host image. The thresholds in each frequency of the host image differ each other. The watermarks where is the same positions are added to the larger coefficients than threshold in the blocks at LL band and middle frequency band in order to prevent the quality deterioration of the host image. The watermarks are inserted in LL band and middle frequency band of the host image. In order to be invisibility of the watermark, the Human Visual System(HVS) is applied to the watermark. We prove the proper embedding method by experiment. We rapidly detect the watermark using this watermarking method. And because the small size watermarks are inserted by HVS, the results confirm the superiority of the proposed method on invisibility and robustness.

An apparatus for the transmission of speech signals reduces the effect of disturbances on the transmission quality. Two microphones are provided for picking up speech signals. Signal processing ensues in three partial frequency bands. The microphone signals are high-pass filtered in a lower frequency band, in the middle frequency band the signal is weighted with a scalar factor, so that this frequency band is damped during the speech pauses, and in the upper frequency band, an adaptive filter is used. At the beginning of the processing, a treble enhancement of the signals ensues, which is canceled by an inverse filter before the output of the improved signal.

Power spectral analysis of fetal heart rate variability has been proposed to provide a non-invasive estimate of autonomic balance. However, there are few systematic data before birth. We therefore examined developmental changes in the frequency power spectrum at very low (0-0.04 Hz), low (0.04-0.15 Hz) and high frequencies (0.15-0.4 Hz), as well as the ratio of low- to high-frequency power (LF/HF), in chronically catheterized, healthy fetal sheep at 0.6 (n = 8), 0.7 (n = 7) and 0.8 gestational age (ga; n = 11). In a second study, 0.8 ga fetuses received either atropine (4.8 mg bolus, then 4.8 mg h(-1) for 30 min, n = 6) or 6-hydroxydopamine (20 mg ml(-1) at 2.5 ml h(-1) for 3 h; n = 9). Data were analysed by sleep state, defined by low-voltage-high-frequency (LV) or high-voltage-low-frequency (HV) EEG. Total spectral power increased with gestational age (P < 0.05), while LF/HF decreased from 0.6 to 0.7 ga. At 0.8 ga, heart rate and LF/HF were significantly higher during HV than LV sleep (P < 0.05). Consistent with this, although total spectral power was not significantly greater during HV sleep, there was a significant interaction between sleep state and frequency band (P = 0.02). Both atropine (P = 0.05) and 6-hydroxydopamine (P < 0.05) were associated with an overall reduction in spectral power but no significant effect on the LF/HF ratio. This study does not support substantial, consistent differences between the frequencies of sympathetic and parasympathetic activity in late-gestation fetal sheep.

Single Watermark Embedding (SWE) technique has been proposed in the [9] by Wong et al. SWE is used to embed a watermark bit sequence into digital images using two secret keys. Another technique used Discrete Cosine Transform (DCT) which enables an 8 × 8 block of an image to be separated into three frequency bands namely the high, middle and low frequency bands. These frequency bands are considered to embed watermark into. It is better to embed watermark in some coefficients in the middle frequency band due to its insensible visibility. In this paper, an improved single watermark embedding (ISWE) is proposed to create a more powerful watermarking scheme that can give better watermarked image quality and enhance watermark’s robustness against JPEG compression attack.

Fragile watermarking is designed to protect content integrity of digital data, meaning to detect any change in the content as well as localize areas that have been changed. In this paper, we propose a fragile watermarking scheme, which embeds dual watermarks (one is naturally fragile and the other is actually robust) in JPEG-2000 compressed images with ROI (region of interest) selection. In this scheme, the fragile watermark is embedded in ROI coefficients at the first decomposition layer, while the robust watermark embedded in ROB (region of backgrounds) coefficients at the third decomposition layer. The combinational working of them is not only capable of detecting malicious changes precisely and flexibly but also estimating the degree of alteration to discriminate intended attacks from unintended ones (e.g., such normal image processing as compression, low-pass filtering, sharpening, etc). Experiments show that our functionally fragile watermarking scheme is scalable in view of detecting attack's intension as well as accuracy in localizing the tampered areas.

The driving motor system of hybrid electric vehicle is under complex vibration conditions, and is subject to the common operating conditions of vehicle and engine. there are some differences between GB/T 18488-2015 and ISO 19453.3:2018 in requirements of working conditions. In this paper, a typical hybrid electrical vehicle (HEV) is used to collect the driving motor vibration condition. The test results show that the vibration energy of the drive motor system of HEV is the most intense in the low frequency band (10Hz $\sim 30$ Hz). It gradually weakens in the middle frequency band (30Hz $\sim 1000$ Hz) and tends to be stable in the high frequency band (1000Hz $\sim 2000$ Hz). However, the vibration intensity at the end of the high frequency band has an increasing trend, which is not reflected in the standards. There may be some deficiency in the vibration resistance assessment of the drive motor system at the high frequency operating point. The determination of vibration energy value is directly related to the acceleration time and acceleration life of the working condition. Compared with other test site specifications, this paper analyzes the reasons why the RMS of the test results is weaker than the standard values. It is proved through the test that the determination of vibration condition needs to determine the reasonable acceleration time of working condition according to the design life of the product, so as to determine the intensity of vibration condition.

This paper presented one novel microstrip antenna with triple-band triple-mode and triple-polarization (TBTMTP) characteristics. In the lower frequency band, the antenna operates in the linearly polarized (LP) dipole mode with omnidirectional radiation pattern along the YOZ plane which is like one horizontally located dipole antenna. In the middle frequency band, the antenna operates in the LP dipole mode with omnidirectional radiation pattern along the XOY plane which is like one vertically located dipole antenna. While in the upper frequency band, the antenna works in the circularly polarized (CP) patch mode with directional radiation patterns along the Z-axis direction. Three modes with different patterns were excited in three different bands because of the mutual-coupling periodical circular patches and inductance metal pins. By etching inverted L shaped slot in the down circular mental patch, CP is realized for the patch mode without affecting the other two modes. The surface current distribution is introduced to analyze the antenna working principle. Due to the compact structure, easy fabrication, radiation pattern selectivity and polarized diversity, this antenna has potential application value in wireless communications in future.

For solving the problems of storage and transmission existing in the high-resolution photogrammetric image, an efficient photogrammetric image encoding method based on Region of Interest (ROI) is proposed according to the application characters of photogrammetric image. The ROIs of the photogrammetric image are automatically detected and extracted using the binarization segmentation and the image dilation algorithm. The description of ROI mask is achieved by combining the coefficient similarity in the same level wavelet sub-bands with specific wavelet filter at an extremely low bit rate. All ROI coefficients are compressed using the lossless method to completely reserve the image information necessary for higher measurement precision. According to the energy concentration features of wavelet transform, only the BG (Back Ground) coefficients belonging to the lowest frequency sub-band are encoded in order to improve total compression ratio. Experimental results show that the ROI mask and the main context information could be described at less than 0.03bit/pixel. Furthermore, the proposed method has improved its compression ratio by more than 6%, and reduced the encoding time above 30% as compared with those of the MaxShift method based on JPGE2000. Thus, the proposed method realizes a high efficiency encoding of photogrammetric image and lays a good foundation for the storage and transmission of the high-resolution photogrammetric image.

A reinforcement-learning-based non-uniform compressed sensing (NCS) framework for time-varying signals is introduced. The proposed scheme, referred to as RL-NCS, aims to boost the performance of signal recovery through an optimal and adaptive distribution of sensing energy among two groups of coefficients of the signal, referred to as region of interest (ROI) coefficients and non-ROI coefficients. The coefficients in ROI usually have greater importance and need to be reconstructed with higher accuracy compared to non-ROI coefficients. In order to accomplish this task, the ROI is predicted at each time-step using two specific approaches. One of these approaches incorporates a long short-term memory (LSTM) network for the prediction. The other approach employs the previous ROI information for predicting the next step ROI. Using the exploration-exploitation technique, a Q-network learns to choose the best approach for designing the measurement matrix. Furthermore, a joint loss function is introduced for the efficient training of the Q-network as well as the LSTM network. The result indicates a significant performance gain for our proposed method, even for rapidly varying signals and reduced number of measurements.

Evidence is presented for local and remote wind‐forcing of deep ocean currents in three frequency bands (corresponding to 3.7–8 day, 8–22 day, and 23–82 day periods) at a two‐year mooring in the eastern North Atlantic. Coherence between current meter velocities and estimated wind stress curl is largest for the deepest instrument (3000 m) and decreases toward the surface. The strongest evidence is found at 3000 m depth in the high frequency (3.7–8 day) band, where both components of velocity are significantly coherent with wind stress curl. Meridional velocity at 3000 m is also significantly coherent with wind stress curl in the low frequency (23–82 day) band, while zonal velocity in the low frequency band and both velocity components in the intermediate frequency (8–22 day) band are only marginally coherent. The results are compared with a stochastically‐forced linear quasi‐geostrophic model that includes a simple representation of the mid‐Atlantic Ridge. Good agreement between predicted and observed coherences is found for the 3000 m currents in the high frequency band, but only partial agreement is found between predicted and observed coherences in the low frequency band, and little agreement in the intermediate band. Kinetic energy levels at 3000 m are well predicted in all three bands. For the shallower instruments (1100 m and 500 m), observed coherences are at most marginally significant and energy levels are considerably higher than the theory predicts. Including an upper‐layer mean flow in the model improves the prediction of energy levels at the shallower instruments.

FV 5 Striato-pallidal connectivity in oscillatory activity of patients with dystonia

To characterize levator ani muscle (LAM) activity in nulligravidas using magnetomyography (MMG) and define MMG characteristics associated with LAM activity with and without accessory muscle contributions. MMG data were collected from eight nulligravidas during rest and voluntary LAM contractions (Kegels) of varying intensity. We utilized simultaneous vaginal manometry and surface electromyography (sEMG) to evaluate for accessory muscle recruitment. Moderate Kegel (MK) MMG trials were sub-selected based on the presence or absence of accessory muscle interaction. Amplitude and spectral-related indicators were calculated across MK epochs: root-mean square (RMS) amplitude, percentage amplitude relative to rest, and relative power spectrum density (rPSD) in three frequency bands (low, middle, high). Ternary diagram characterized rPSD from selected Kegels and ROC analysis was performed to identify cut-points to differentiate MKs from interacting MKs. Nineteen MMG recordings were obtained. Amplitude and spectral parameters were significantly different between isolated and interacting MK epochs. Mean RMS and power values of the isolated MK were, respectively, 120.66 ± 43.8 fT and 1.72 ± 1.44 (T2 /Hz)*10-28 . Amplitudes of MK were 64% and 117 higher than baseline activities for the isolated and interacting epochs, respectively. ROC curves reveled cut-off points on low and middle frequency bands that achieved perfect separation (ROC-AUC = 1.0) between isolated and interacting MK. Our study demonstrates that MMG, a novel biomagnetic technique, allows precise detection and characterization of normal female pelvic floor function. Results show that isolated moderate voluntary contraction of the LAMs produces distinct MMG amplitude and spectral characteristics compared with Kegels involving co-activation of other muscle groups.

In this work, we propose a data embedding scheme in MPEG-1/Audio Layer II compressed domain. Data embedding is conducted every AAU by using side information (location of sub-band allocated audio signal) as a data carrier. In general, non-zero signals concentrates in low and middle frequency bands. Therefore we utilize sub-bands that are not allocated audio signal in high frequency bands to embed information. The proposed scheme can increase payload while achieving rewritable (reversible) data embedding by choosing appropriate parameter. We verify the basic performance of our scheme through computer simulation by using some voice and music signals. 1.

A compact four-element dual-band multiple-input and multiple-output (MIMO) antenna system is proposed to achieve high isolation and low channel capacity loss. The MIMO antenna was designed and optimized to cover the dual-frequency bands; the first frequency band is a wide band, and it covers the frequency range of 1550–2650 MHz, while the other frequency band covers the 3350–3650 MHz range. The measured wide-band impedance bandwidths of 1.1 GHz and 300 MHz were achieved in the lower and upper frequency bands, respectively. The proposed structure consists of four novel antenna elements, along with a plus-sign-shaped ground structure on an FR4 substrate. The overall electrical size of the whole dual-band MIMO antenna system is (W) × (L) × (H) for the lower frequency band. It achieved greater than 10 and 19 dB isolation in the lower and upper frequency bands, respectively. The antenna system accomplished an envelope correlation coefficient of in the lower frequency band, while it achieved in the higher frequency band. The computed channel capacity loss remained less than almost 0.4 bits/s/Hz in both frequency bands. Therefore, it achieved good performance in both frequency bands, with the additional advantage of a compact size. The proposed MIMO antenna is suitable for compact handheld devices and smartphones used for GSM (Global System for Mobiles), UMTS (Universal Mobile Telecommunications Service), WCDMA (Wideband Code Division Multiple Access), LTE (Long Term Evolution), 5G sub-6 GHz, PCS (Personal Communications Service), and WLAN (wireless local area network) applications.

In spreading spectrum audio watermarking system, the random noise on full-frequency is very common, which greatly affects the receiving efficiency. In this paper, we proposed an optimized receiving method, which is specially used to resist full-frequency domain random noise. For most audio signals, especially voice signals, the energy is mainly concentrated in the middle and low frequency bands, and the energy in high frequency band is less. We take advantage of this feature to optimize the performance of spreading spectrum audio watermarking system. Experimental results show that our optimized receiving method will ultimately improve the decoding performance.