Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9980134
Y. Sato, S. Maeda, M. Akasaka, M. Nishigaki, Tetsushi Ohki
Facial images contain sensitive attributes such as skin color, and the elimination of them from the input in the face recognition is not easy. In addition, the input data includes the influence of the environment in which the system is actually used, so the interaction between sensitive attributes and the environment may make it inherently difficult for the facial feature extractor to extract facial features. Therefore, studies on the fairness of face recognition should consider the fairness of environmental factors. Common datasets used to evaluate the fairness of face recognition includes a variety of environmental factors, and the fairness evaluated by these datasets are usually the fairness in a typical shooting environment. However, a dataset that includes only extremely biased environmental factors potentially results in less equity among attributes. We construct a dataset with pseudo-biased environmental factors by dynamically changing environmental factors such as brightness in the test data. The results also show that the biased environmental factors deteriorate the fairness inter-attribute. Also, we showed that the distinguished attributes in terms of fairness in a biased environment vary based on the architecture of the model and the training dataset.
{"title":"A Fair Model is not Fair in a Biased Environment","authors":"Y. Sato, S. Maeda, M. Akasaka, M. Nishigaki, Tetsushi Ohki","doi":"10.23919/APSIPAASC55919.2022.9980134","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9980134","url":null,"abstract":"Facial images contain sensitive attributes such as skin color, and the elimination of them from the input in the face recognition is not easy. In addition, the input data includes the influence of the environment in which the system is actually used, so the interaction between sensitive attributes and the environment may make it inherently difficult for the facial feature extractor to extract facial features. Therefore, studies on the fairness of face recognition should consider the fairness of environmental factors. Common datasets used to evaluate the fairness of face recognition includes a variety of environmental factors, and the fairness evaluated by these datasets are usually the fairness in a typical shooting environment. However, a dataset that includes only extremely biased environmental factors potentially results in less equity among attributes. We construct a dataset with pseudo-biased environmental factors by dynamically changing environmental factors such as brightness in the test data. The results also show that the biased environmental factors deteriorate the fairness inter-attribute. Also, we showed that the distinguished attributes in terms of fairness in a biased environment vary based on the architecture of the model and the training dataset.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"54 35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123806519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9979898
Qi Song, Zebin Peng, Luchen Ji, Xiaochen Yang, Xiaoxu Li
Deep neural networks have outperformed humans on some image recognition and classification tasks. However, with the emergence of various novel classes, it remains a chal-lenge to continuously expand the learning capability of such networks from a limited number of labeled samples. Metric-based approaches have been playing a key role in few-shot image classification, but most of them measure the distance between samples in the metric space using only a single metric function. In this paper, we propose a Dual Prototypical Network (DPN) to improve the test-time robustness of the classical prototypical network. The proposed method not only focuses on the distance of the original features, but also adds perturbation noise to the image and calculates the distance of noisy features. By enforcing the model to predict well under both metrics, more representative and robust class prototypes are learned and thus lead to better generalization performance. We validate our method on three fine-grained datasets in both clean and noisy settings.
{"title":"Dual Prototypical Network for Robust Few-shot Image Classification","authors":"Qi Song, Zebin Peng, Luchen Ji, Xiaochen Yang, Xiaoxu Li","doi":"10.23919/APSIPAASC55919.2022.9979898","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9979898","url":null,"abstract":"Deep neural networks have outperformed humans on some image recognition and classification tasks. However, with the emergence of various novel classes, it remains a chal-lenge to continuously expand the learning capability of such networks from a limited number of labeled samples. Metric-based approaches have been playing a key role in few-shot image classification, but most of them measure the distance between samples in the metric space using only a single metric function. In this paper, we propose a Dual Prototypical Network (DPN) to improve the test-time robustness of the classical prototypical network. The proposed method not only focuses on the distance of the original features, but also adds perturbation noise to the image and calculates the distance of noisy features. By enforcing the model to predict well under both metrics, more representative and robust class prototypes are learned and thus lead to better generalization performance. We validate our method on three fine-grained datasets in both clean and noisy settings.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122603981","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9979994
Jing-Ming Guo, Sankarasrinivasan Seshathiri
As most multimedia is compressed for optimal storage or transmission, the watermarking during compression is more appropriate and demanding. Multi-tone block truncation coding image (MT-BTC) is a latest and superior version of halftone based block truncation coding images. In this paper, a novel watermarking strategy is proposed for MT-BTC images using the adaptive dither array selection (ADAS) and inter-tone shifting. ADAS method utilize dither array constructed using two different gaussian filters based on the Human Visual System (HVS) model to embed the watermark information. Further, various configuration of dither array such as actual, conjugate, transpose and transpose conjugate is used to embed more data. Moreover, a new approach termed inter-tone shifting is also proposed to improve the decoding rate and security. The decoding scheme is performed using the pattern similarity on the dithered watermarked image. For result validation, a 2,4,6 and 8-Tone watermark image is embedded in the 4-Tone MTBTC image. From extensive analysis on decoding rate and robustness, it has been validated that the proposed scheme outperforms many halftone watermarking and consistent with conventional BTC methods.
由于大多数多媒体都是为了优化存储或传输而进行压缩的，因此压缩过程中的水印更加合适，要求也更高。多色调分块截断编码图像(MT-BTC)是基于半色调分块截断编码图像的最新、优越版本。本文提出了一种基于自适应抖动阵列选择(ADAS)和色调间移位的MT-BTC图像水印策略。ADAS方法利用基于人类视觉系统(HVS)模型的两种不同高斯滤波器构建的抖动阵列来嵌入水印信息。此外，还采用了实际抖动阵列、共轭抖动阵列、转置抖动阵列和转置共轭抖动阵列等多种构型来嵌入更多的数据。此外，为了提高译码率和安全性，还提出了一种新的译码方法——音间移位。利用抖动水印图像的模式相似度进行解码。为了验证结果，在4- tone mbittc图像中嵌入2,4,6和8-Tone水印图像。通过对译码率和鲁棒性的广泛分析，验证了该方案优于许多半色调水印，与传统的BTC方法一致。
{"title":"A Unified Compression and Watermarking Scheme for MT-BTC Images","authors":"Jing-Ming Guo, Sankarasrinivasan Seshathiri","doi":"10.23919/APSIPAASC55919.2022.9979994","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9979994","url":null,"abstract":"As most multimedia is compressed for optimal storage or transmission, the watermarking during compression is more appropriate and demanding. Multi-tone block truncation coding image (MT-BTC) is a latest and superior version of halftone based block truncation coding images. In this paper, a novel watermarking strategy is proposed for MT-BTC images using the adaptive dither array selection (ADAS) and inter-tone shifting. ADAS method utilize dither array constructed using two different gaussian filters based on the Human Visual System (HVS) model to embed the watermark information. Further, various configuration of dither array such as actual, conjugate, transpose and transpose conjugate is used to embed more data. Moreover, a new approach termed inter-tone shifting is also proposed to improve the decoding rate and security. The decoding scheme is performed using the pattern similarity on the dithered watermarked image. For result validation, a 2,4,6 and 8-Tone watermark image is embedded in the 4-Tone MTBTC image. From extensive analysis on decoding rate and robustness, it has been validated that the proposed scheme outperforms many halftone watermarking and consistent with conventional BTC methods.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122884414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9979829
Yitong Chen, Wen Zhang
Due to its satisfactory performance and no need for room impulse response information, the adaptive weighted prediction error (AWPE) algorithm is promising for speech dereverberation in practice. However, the robustness of AWPE to additive noise is low. To alleviate this problem, this paper proposes a variant of the AWPE algorithm that is based on eigen-decomposition of the signal auto-correlation matrix to construct the reference signal. By using the dominant eigenvector as the reference signal, a linear prediction filter is designed which has a better performance to predict the late reverberation even when the additive noise level is high. To reduce the computational complexity of the standard eigen-decomposition operation in the proposed AWPE variant, an online eigenvector extraction algorithm based on a fixed-point iteration algorithm is presented. Simulations are conducted to validate the effectiveness and robustness of the proposed algorithms over the standard AWPE algorithm.
{"title":"Robust Speech Dereverberation Based on Adaptive Weighted Prediction Error Algorithm with Eigenvector Extraction","authors":"Yitong Chen, Wen Zhang","doi":"10.23919/APSIPAASC55919.2022.9979829","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9979829","url":null,"abstract":"Due to its satisfactory performance and no need for room impulse response information, the adaptive weighted prediction error (AWPE) algorithm is promising for speech dereverberation in practice. However, the robustness of AWPE to additive noise is low. To alleviate this problem, this paper proposes a variant of the AWPE algorithm that is based on eigen-decomposition of the signal auto-correlation matrix to construct the reference signal. By using the dominant eigenvector as the reference signal, a linear prediction filter is designed which has a better performance to predict the late reverberation even when the additive noise level is high. To reduce the computational complexity of the standard eigen-decomposition operation in the proposed AWPE variant, an online eigenvector extraction algorithm based on a fixed-point iteration algorithm is presented. Simulations are conducted to validate the effectiveness and robustness of the proposed algorithms over the standard AWPE algorithm.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121241582","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9979983
Kah-Meng Cheong, Yih-Liang Shen, T. Chi
Psychoacoustic active noise control (PANC) systems have been proposed to improve the noise reduction performance of active noise control (ANC) systems by considering hearing properties. PANC systems are usually implemented in the filtered@ $mathbf{x}$ least mean square (FxLMS) architecture. In this paper, we propose a PANC system using a simple and stable fast affine projection (FAP) algorithm in the filtered-x architecture, namely the filtered-x conjugate gradient fast affine projection (FxCGFAP) PANC. The proposed FxCGFAP PANC system converges faster than typical FxLMS ANC and PANC systems. The proposed PANC system considers not only the sensitivity of human hearing to frequency but also the sensitivity to spectral modulation. Objective and subjective evaluations have been conducted. The evaluation results show that the proposed system outperforms the other three systems in terms of objective loudness scores and subjective ratings. The proposed system has been implemented on the Tensilica HiFi3 DSP platform with the fixed-point data format.
{"title":"A Fast Converge Spectral Modulation Sensitive Active Noise Control System","authors":"Kah-Meng Cheong, Yih-Liang Shen, T. Chi","doi":"10.23919/APSIPAASC55919.2022.9979983","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9979983","url":null,"abstract":"Psychoacoustic active noise control (PANC) systems have been proposed to improve the noise reduction performance of active noise control (ANC) systems by considering hearing properties. PANC systems are usually implemented in the filtered@ $mathbf{x}$ least mean square (FxLMS) architecture. In this paper, we propose a PANC system using a simple and stable fast affine projection (FAP) algorithm in the filtered-x architecture, namely the filtered-x conjugate gradient fast affine projection (FxCGFAP) PANC. The proposed FxCGFAP PANC system converges faster than typical FxLMS ANC and PANC systems. The proposed PANC system considers not only the sensitivity of human hearing to frequency but also the sensitivity to spectral modulation. Objective and subjective evaluations have been conducted. The evaluation results show that the proposed system outperforms the other three systems in terms of objective loudness scores and subjective ratings. The proposed system has been implemented on the Tensilica HiFi3 DSP platform with the fixed-point data format.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127731112","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9979875
Qilin Zhang, S. Ibi, Takumi Takahashi, H. Iwai
This paper proposes a deep unfolding-aided sum-product algorithm (SPA) for error correction decoding of cyclic redundancy check (CRC) coded short message. SPA is a practical decoding algorithm for linear codes without requiring enormous computational complexity. However, if the SPA is used as it is for CRC codes, belief correlation and outliers will be induced in the iterative decoding process, resulting in lousy correction capability. To compensate for this drawback, we design a SPA-based decoding process for CRC code that incorporates a data-driven design based on deep learning and learning optimization of in-ternal trainable parameters. Considering the operation principle of soft-decision decoder, a novel loss function based on a weighted average of negentropy, which is a key measure to evaluate the Gaussianity, and BCE of the decoder output is proposed. Numerical results show that the proposed algorithm improves the bit error rate (BER) performance with deep unfolding and negentropy-aware loss function.
{"title":"Deep Unfolding-Aided Sum-Product Algorithm for Error Correction of CRC Coded Short Message","authors":"Qilin Zhang, S. Ibi, Takumi Takahashi, H. Iwai","doi":"10.23919/APSIPAASC55919.2022.9979875","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9979875","url":null,"abstract":"This paper proposes a deep unfolding-aided sum-product algorithm (SPA) for error correction decoding of cyclic redundancy check (CRC) coded short message. SPA is a practical decoding algorithm for linear codes without requiring enormous computational complexity. However, if the SPA is used as it is for CRC codes, belief correlation and outliers will be induced in the iterative decoding process, resulting in lousy correction capability. To compensate for this drawback, we design a SPA-based decoding process for CRC code that incorporates a data-driven design based on deep learning and learning optimization of in-ternal trainable parameters. Considering the operation principle of soft-decision decoder, a novel loss function based on a weighted average of negentropy, which is a key measure to evaluate the Gaussianity, and BCE of the decoder output is proposed. Numerical results show that the proposed algorithm improves the bit error rate (BER) performance with deep unfolding and negentropy-aware loss function.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"342 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132632371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9980284
Hiromu Takayama, Tatsuya Yokota
Recently, signal completion methods using delay-embedding transforms (DT) have been actively studied. Since the DT is an operation to transform a signal into a Hankel matrix, the high computational cost associated with the increase in data size is an issue. In this study, we consider modeling smooth signals based on inverse delay-embedding instead of delay-embedding. We propose a new algorithm that incorporates the properties of the delay-embedding-based methods while reducing the computational cost. The proposed algorithm takes advantage of the inverse delay-embedding being a cyclic convolution, and the computational complexity can be reduced to $mathcal{O}(NlogN)$ by transforming the optimization problem to Fourier space. Numerical experiments with typical signals and audio data show the effectiveness of the proposed algorithm in signal declipping and completion problems.
{"title":"Fast Signal Completion Algorithm with Cyclic Convolutional Smoothing","authors":"Hiromu Takayama, Tatsuya Yokota","doi":"10.23919/APSIPAASC55919.2022.9980284","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9980284","url":null,"abstract":"Recently, signal completion methods using delay-embedding transforms (DT) have been actively studied. Since the DT is an operation to transform a signal into a Hankel matrix, the high computational cost associated with the increase in data size is an issue. In this study, we consider modeling smooth signals based on inverse delay-embedding instead of delay-embedding. We propose a new algorithm that incorporates the properties of the delay-embedding-based methods while reducing the computational cost. The proposed algorithm takes advantage of the inverse delay-embedding being a cyclic convolution, and the computational complexity can be reduced to $mathcal{O}(NlogN)$ by transforming the optimization problem to Fourier space. Numerical experiments with typical signals and audio data show the effectiveness of the proposed algorithm in signal declipping and completion problems.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133511797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9980058
Amir Mohammadisrab, Poorya Aghaomidi, Jalil Mazloum, M. Akbarzadeh, M. Orooji, N. Mokari, H. Yanikomeroglu
In this paper, we study the performance of a deep adaptive denoising auto-encoder network (DeepADAENet) for electrocardiogram (ECG) signal noise cancelation in the time-frequency domain for practical use cases. In order to achieve a higher resolution in distinguishing the noise from valuable data, the fractional Stockwell transform (FrST) is exploited to convert the ECG to the time-frequency image. The magnitude of the time-frequency version of the ECG is noise-canceled using DeepADAENet. Then, inverse FrST is utilized to return the denoised time-frequency ECG into the time domain. Furthermore, we use the MIT-BID Apnea-ECG database (APNEA-ECG) for preparing the dataset due to various physiologies and records compared with other ECG databases. Moreover, muscle artifacts (MA), baseline wander (BW), and electrode motion (EM) from the MIT-BID Noise Stress Test Database (NSTDB) are utilized to make noisy this clean dataset. The ECG signals recorded by non-clinical devices contain more noise than clinical recording. Accordingly, by changing the coefficient and frequency of noise resources, we attempt to close the simulated noisy signal to reality. Results reveal the excellent performance of DeepADAENet compared with similar work in terms of signal-to-noise ratio (SNR), root mean square error (RMSE), and percent root mean square difference (PRD).
{"title":"Deep Adaptive Denoising Auto-Encoder Networks for ECG Noise Cancellation via Time-Frequency Domain","authors":"Amir Mohammadisrab, Poorya Aghaomidi, Jalil Mazloum, M. Akbarzadeh, M. Orooji, N. Mokari, H. Yanikomeroglu","doi":"10.23919/APSIPAASC55919.2022.9980058","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9980058","url":null,"abstract":"In this paper, we study the performance of a deep adaptive denoising auto-encoder network (DeepADAENet) for electrocardiogram (ECG) signal noise cancelation in the time-frequency domain for practical use cases. In order to achieve a higher resolution in distinguishing the noise from valuable data, the fractional Stockwell transform (FrST) is exploited to convert the ECG to the time-frequency image. The magnitude of the time-frequency version of the ECG is noise-canceled using DeepADAENet. Then, inverse FrST is utilized to return the denoised time-frequency ECG into the time domain. Furthermore, we use the MIT-BID Apnea-ECG database (APNEA-ECG) for preparing the dataset due to various physiologies and records compared with other ECG databases. Moreover, muscle artifacts (MA), baseline wander (BW), and electrode motion (EM) from the MIT-BID Noise Stress Test Database (NSTDB) are utilized to make noisy this clean dataset. The ECG signals recorded by non-clinical devices contain more noise than clinical recording. Accordingly, by changing the coefficient and frequency of noise resources, we attempt to close the simulated noisy signal to reality. Results reveal the excellent performance of DeepADAENet compared with similar work in terms of signal-to-noise ratio (SNR), root mean square error (RMSE), and percent root mean square difference (PRD).","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134038951","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9980026
An Gia Vien, Truong Thanh Nhat Mai, Seonghyun Park, Gahyeong Kim, Chul Lee
We propose a high dynamic range (HDR) image synthesis algorithm based on enhanced bidirectional motion estimation using feature refinement. First, we extract multiscale features from input low dynamic range (LDR) images and then estimate accurate motion vector fields between them in a coarse-to-fine manner via progressive refinement. Then, we estimate adaptive local kernels to merge only valid information in the spatio-exposed neighboring pixels for synthesis. Finally, we refine the initially merged image by exploiting global information to further improve synthesis performance. Experimental results show that the proposed algorithm outperforms state-of-the-art algorithms in quantitative and qualitative comparisons.
{"title":"Enhanced Bidirectional Motion Estimation Using Feature Refinement for HDR Imaging","authors":"An Gia Vien, Truong Thanh Nhat Mai, Seonghyun Park, Gahyeong Kim, Chul Lee","doi":"10.23919/APSIPAASC55919.2022.9980026","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9980026","url":null,"abstract":"We propose a high dynamic range (HDR) image synthesis algorithm based on enhanced bidirectional motion estimation using feature refinement. First, we extract multiscale features from input low dynamic range (LDR) images and then estimate accurate motion vector fields between them in a coarse-to-fine manner via progressive refinement. Then, we estimate adaptive local kernels to merge only valid information in the spatio-exposed neighboring pixels for synthesis. Finally, we refine the initially merged image by exploiting global information to further improve synthesis performance. Experimental results show that the proposed algorithm outperforms state-of-the-art algorithms in quantitative and qualitative comparisons.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133801581","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-07DOI: 10.23919/APSIPAASC55919.2022.9980262
Long Ngo, Q. Pham
In this research, we proposed a new digital coherent holographic configuration for accurately measuring the three-dimensional (3D) vibration of the object. The vibration was indirectly measured by the displacement of the three mirrors attached on the object. The hologram recorded by the camera consisting of 6 sub-holograms can be separated by Fourier transform and appropriated spatial band-pass filters. Three phase sets extracted from 3 sub-holograms of the reference mirror and 3 object mirrors were used to calculate the displacement of the object in 3D directions. The relation between the displacement of the object and the phases of the sub-holograms was related to the wavelength of the light source, therefore this allows observing the vibration of the object with nano-scale accuracy in z direction and much smaller than the pixel size of the camera accuracy in x and y directions.
{"title":"Vibration measurement using spatial shifting coherent digital holography","authors":"Long Ngo, Q. Pham","doi":"10.23919/APSIPAASC55919.2022.9980262","DOIUrl":"https://doi.org/10.23919/APSIPAASC55919.2022.9980262","url":null,"abstract":"In this research, we proposed a new digital coherent holographic configuration for accurately measuring the three-dimensional (3D) vibration of the object. The vibration was indirectly measured by the displacement of the three mirrors attached on the object. The hologram recorded by the camera consisting of 6 sub-holograms can be separated by Fourier transform and appropriated spatial band-pass filters. Three phase sets extracted from 3 sub-holograms of the reference mirror and 3 object mirrors were used to calculate the displacement of the object in 3D directions. The relation between the displacement of the object and the phases of the sub-holograms was related to the wavelength of the light source, therefore this allows observing the vibration of the object with nano-scale accuracy in z direction and much smaller than the pixel size of the camera accuracy in x and y directions.","PeriodicalId":382967,"journal":{"name":"2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133853087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：