• The Journal of the Acoustical Society of Korea
• /
• v.42 no.6
• /
• pp.544-551
• /
• 2023

Speech enhancement used to improve the perceptual quality and intelligibility of noise speech has been studied as a method using a complex-valued spectrum that can improve both magnitude and phase in a method using a magnitude spectrum. In this paper, a study was conducted on how to apply attention mechanism to complex-valued spectrum-based speech enhancement systems to further improve the intelligibility and quality of noise speech. The attention is performed based on additive attention and allows the attention weight to be calculated in consideration of the complex-valued spectrum. In addition, the global average pooling was used to consider the importance of the feature map. Complex-valued spectrum-based speech enhancement was performed based on the Deep Complex U-Net (DCUNET) model, and additive attention was conducted based on the proposed method in the Attention U-Net model. The results of the experiments on noise speech in a living room environment showed that the proposed method is improved performance over the baseline model according to evaluation metrics such as Source to Distortion Ratio (SDR), Perceptual Evaluation of Speech Quality (PESQ), and Short Time Object Intelligence (STOI), and consistently improved performance across various background noise environments and low Signal-to-Noise Ratio (SNR) conditions. Through this, the proposed speech enhancement system demonstrated its effectiveness in improving the intelligibility and quality of noisy speech.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2004-2007
• /
• 2003

This paper describes a method of performance enhancement using Flatness Mesure(FM) for the Gaussian Mixture Model(GMM) face recognition systems. Using this measure we discard the frames having low information before training and test. As the result, the performance increases about 9％ in the lower mixtures and calculation burden is decreased. As well, the recognition error rate is decreased under the illumination change surroundings. We use the 2D DCT coefficients lot face feature vectors and experiments are carried out on the Olivetti Research Laboratory (ORL) face database.

• The KIPS Transactions:PartB
• /
• v.19B no.2
• /
• pp.113-122
• /
• 2012

This paper presents a new stereo matching technique which is based on the matching of feature windows and feature links. The proposed method uses the FAST feature detector to find image features in stereo images and determines the correspondences of the detected features in the stereo images. We define a feature window which is an image region containing several image features. The proposed technique consists of two matching steps. First, a feature window is defined in a standard image and its correspondence is found in a reference image. Second, the corresponding features between the matched windows are determined by using the feature link technique. If there is no correspondence for an image feature in the standard image, it's disparity is interpolated by neighboring feature sets. We evaluate the accuracy of the proposed technique by comparing our results with the ground truth of in a stereo image database. We also compare the matching accuracy and computation time with two conventional feature-based stereo matching techniques.

• Phonetics and Speech Sciences
• /
• v.5 no.1
• /
• pp.11-16
• /
• 2013

This paper proposes robust feature extraction for accurate voice activity detection (VAD). VAD is one of the principal modules for speech signal processing such as speech codec, speech enhancement, and speech recognition. Noisy environments contain nonstationary noises causing the accuracy of the VAD to drastically decline because the fluctuation of features in the noise intervals results in increased false alarm rates. In this paper, in order to improve the VAD performance, harmonic-weighted energy is proposed. This feature extraction method focuses on voiced speech intervals and weighted harmonic-to-noise ratios to determine the amount of the harmonicity to frame energy. For performance evaluation, the receiver operating characteristic curves and equal error rate are measured.

• Journal of Korea Multimedia Society
• /
• v.21 no.2
• /
• pp.108-116
• /
• 2018

In this paper, we propose an image enhancement method based on Multi-Scale Retinex theory that designs Unsharp Masking Filter (UMF) and emphasizes the contrast ratio adaptively. Unsharp Masking (UM) technique emphasizes image sharpness and improves contrast ratio by adding high frequency component to the original image. The high frequency component is obtained by differentiating between original image and low frequency image. In this paper, we present how to design an UMF kernel and to adaptively apply it to increase the contrast ratio according to multi-scale retinex theory which resembles human visual system. Experimental results show that the proposed method has better quantitative performance indexes such as PSNR, ambe & SSIM and better qualitative feature like halo artifact suppression.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.326-327
• /
• 2007

We investigated whether the CT images of hepatic lesions could be analyzed by computer-aided diagnosis (CAD) tool. We retrospectively reanalyzed 14 liver CT images (10 hepatocellular cancers and 4 benign liver lesions; patients who presented with hepatic masses). The hepatic lesions on CT were segmented by rectangular ROI technique and the morphologic features were extracted and quantitated using fractal texture analysis. The contrast enhancement of hepatic lesions was also quantified and added to the differential diagnosis. The best discriminating function combining the textural features and the values of contrast enhancement of the lesions was created using linear discriminant analysis. Textural feature analysis showed moderate accuracy in the differential diagnosis of hepatic lesions, but statistically insignificant. Combining textural analysis and contrast enhancement value resulted in improved diagnostic accuracy, but further studies are needed.

• 한국방재학회:학술대회논문집
• /
• 2011.02a
• /
• pp.39-39
• /
• 2011

This study presents an effective vision-based system for cable bridge damage detection. In theory, cable bridges need to be inspected the outer as well as the inner part. Starting from August 2010, a new research project supported by Korea Ministry of Land, Transportation Maritime Affairs(MLTM) was initiated focusing on the damage detection of cable system. In this study, only the surface damage detection algorithm based on a vision-based system will be focused on, an overview of the vision-based cable damage detection is given in Fig. 1. Basically, the algorithm combines the image enhancement technique with principal component analysis(PCA) to detect damage on cable surfaces. In more detail, the input image from a camera is processed with image enhancement technique to improve image quality, and then it is projected into PCA sub-space. Finally, the Mahalanobis square distance is used for pattern recognition. The algorithm was verified through laboratory tests on three types of cable surface. The algorithm gave very good results, and the next step of this study is to implement the algorithm for real cable bridges.

• Journal of Information Processing Systems
• /
• v.19 no.5
• /
• pp.663-672
• /
• 2023

Most vehicle detection methods have poor vehicle feature extraction performance at night, and their robustness is reduced; hence, this study proposes a night vehicle detection method based on style transfer image enhancement. First, a style transfer model is constructed using cycle generative adversarial networks (cycleGANs). The daytime data in the BDD100K dataset were converted into nighttime data to form a style dataset. The dataset was then divided using its labels. Finally, based on a YOLOv5s network, a nighttime vehicle image is detected for the reliable recognition of vehicle information in a complex environment. The experimental results of the proposed method based on the BDD100K dataset show that the transferred night vehicle images are clear and meet the requirements. The precision, recall, mAP@.5, and mAP@.5:.95 reached 0.696, 0.292, 0.761, and 0.454, respectively.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.4
• /
• pp.687-692
• /
• 2006

In a conventional ICA(Independent Component Analysis) based face recognition method, PCA(Principal Component Analysis) first is used for feature extraction, ICA learning method then is applied for feature enhancement in the reduced dimension. It is not considered that a necessary component can be located in the discarded feature space. In the new ICA(NICA), learning extracts features using the magnitude of kurtosis (4-th order central moment or cumulant). But, the pure ICA method can not discard noise effectively. The synergy effect of PCA and ICA can be achieved if PCA is used for noise reduction filter. Namely, PCA does whitening and noise filtering. ICA performs feature extraction. Experiment results show the effectiveness of the new ICA method compared to the conventional ICA approach.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.37 no.8
• /
• pp.759-766
• /
• 2009

The purpose of this paper is an image processing algorithm development as a base research achieving performance enhancement of integrated navigation system. We used the SIFT (Scale Invariant Feature Transform) algorithm for image processing, and developed feature point matching filter for rejecting mismatched points. By applying the proposed algorithm, it is obtained better result than other methods of parameter tuning and KLT based feature point tracking. For further study, integration with INS and algorithm optimization for the real-time implementation are under investigation.