研究者詳細 - 中澤　篤志

論文 - 中澤　篤志

分割表示 105 件中 21 - 40 件目／全件表示 >>

Evaluating imitation and rule-based behaviors of eye contact and blinking using an android for conversation

Akishige Yuguchi, Tetsuya Sano, Gustavo Alfonso Garcia Ricardez, Jun Takamatsu, Atsushi Nakazawa, Tsukasa Ogasawara

ADVANCED ROBOTICS 35 ( 15 ) 907 - 918 2021年5月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：TAYLOR & FRANCIS LTD

In this paper, we investigate which approach to generate eye behaviors using an android robot makes what impressions on humans and clarify which are the important factors for attractive eye behaviors. Thus, we evaluate the human impression of eye behaviors displayed by an android robot while talking to a human by comparing the motions generated by the two approaches. In the first approach, we develop a method to imitate human eye behavior obtained from eye trackers. In the second approach, we control the eye direction, eye-contact duration, and eyeblinks based on the findings of human eye behavior in psychology and cognitive research. In the experiments, we asked male and female subjects to evaluate their impression by comparing the eye behaviors with an android that controls the eye-contact duration and eyeblinks by editing the imitation parameters or the rule-based behavior. In the experiments, we asked subjects to evaluate their impression of different eye behaviors displayed by an android. The eye behaviors were generated by modifying the imitation parameters or the rule-based behavior, which resulted in different eye-contact duration and eyeblink duration and timing. From the results, we found that (1) the imitation and rule-based behaviors showed no difference in terms of human-likeness, (2) the 3-second eye contact obtained better scores regardless of the imitation or rule-based eye behavior, (3) the subjects might regard the long eyeblinks as voluntary eyeblinks, with the intention to break eye contact, and (4) female subjects preferred short eyeblinks rather than long ones and considered that short eyeblinks might be one of the keys to making eye contact more suitable, in contrast to male subjects who preferred long eyeblinks.

DOI： 10.1080/01691864.2021.1928544

他者の視線方向・表情と参加者の体勢が介護ベッド上での対人認知に及ぼす影響-ヘッドマウントディスプレイを使用した検討-

布井雅人, 吉川左紀子, 中澤篤志

電子情報通信学会論文誌 A(Web) J104-A ( 2 ) 2021年

Visual Place Recognition from Eye Reflection

Yuki Ohshima, Kyosuke Maeda, Yusuke Edamoto, Atsushi Nakazawa

IEEE Access 9 57364 - 57371 2021年

掲載種別：研究論文（学術雑誌）

The cornea in the human eye reflects incoming environmental light, which means we can obtain information about the surrounding environment from the corneal reflection in facial images. In recent years, as the quality of consumer cameras increases, this has caused privacy concerns in terms of identifying the people around the subject or where the photo is taken. This paper investigates the security risk of eye corneal reflection images: specifically, visual place recognition from eye reflection images. First, we constructed two datasets containing pairs of scene and corneal reflection images. The first dataset is taken in a virtual environment. We showed pre-captured scene images in a 180-degree surrounding display system and took corneal reflections from subjects. The second dataset is taken in an outdoor environment. We developed several visual place recognition algorithms, including CNN-based image descriptors featuring a naive Siamese network and AFD-Net combined with entire image feature representations including VLAD and NetVLAD, and compared the results. We found that AFD-Net+VLAD performed the best and was able to accurately determine the scene in 73.08% of the top-five candidate scenes. These results demonstrate the potential to estimate the location at which a facial picture was taken, which simultaneously leads to a) positive applications such as the localization of a robot while conversing with persons and b) negative scenarios including the security risk of uploading facial images to the public.

DOI： 10.1109/ACCESS.2021.3071406

GAN-based Style Transformation to Improve Gesture-recognition Accuracy

Noeru Suzuki, Yuki Watanabe, Atsushi Nakazawa

Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 4 ( 4 ) 154 - 20 2020年12月

掲載種別：研究論文（学術雑誌）

Gesture recognition and human-Activity recognition from multi-channel sensory data are important tasks in wearable and ubiquitous computing. In these tasks, increasing both the number of recognizable activity classes and the recognition accuracy is essential. However, this is usually an ill-posed problem because individual differences in the same gesture class may affect the discrimination of different gesture classes. One promising solution is to use personal classifiers, but this requires personal gesture samples for re-Training the classifiers. We propose a method of solving this issue that obtains personal gesture classifiers using few user gesture samples, thus, achieving accurate gesture recognition for an increased number of gesture classes without requiring extensive user calibration the novelty of our method is introducing a generative adversarial network (GAN)-based style transformer to 'generate' a user's gesture data the method synthesizes the gesture examples of the target class of a target user by transforming of a) gesture data into another class of the same user (intra-user transformation) or b) gesture data of the same class of another user (inter-user transformation) the synthesized data are then used to train the personal gesture classifier. We conducted comprehensive experiments using 1) different classifiers including SVM and CNN, 2) intra-and inter-user transformations, 3) various data-missing patterns, and 4) two different types of sensory data. Results showed that the proposed method had an increased performance. Specially, the CNN-based classifiers increased in average accuracy from 0.747 to 0.822 in the CheekInput dataset and from 0.856 to 0.899 in the USC-HAD dataset. Moreover, the experimental results with various data-missing conditions revealed a relation between the number of missing gesture classes and the accuracy of the existing and proposed methods, and we were able to clarify several advantages of the proposed method these results indicate the potential of considerably reducing the number of required training samples of target users.

DOI： 10.1145/3432199

A Generative Self-Ensemble Approach to Simulated+Unsupervised Learning

Yu Mitsuzum, Go Irie, Akisato Kimura, Atsushi Nakazawa

Proceedings - International Conference on Image Processing, ICIP 2020-October 2151 - 2155 2020年10月

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：IEEE

In this paper, we consider Simulated and Unsupervised (S+U) learning which is a problem of learning from labeled synthetic and unlabeled real images. After translating the synthetic images to real ones, existing S+U learning methods use only the labeled synthetic images for training a predictor (e.g., a regression function) and ignore the target real images, which may result in unsatisfactory prediction performance. Our approach utilizes both synthetic and real images to train the predictor. The main idea of ours is to involve a self-ensemble learning framework into S+U learning. More specifically, we require the prediction results for an unlabeled real image to be consistent between 'teacher' and 'student' predictors, even after some perturbations are added to the image. Furthermore, aiming at generating diverse perturbations along the underlying data manifold, we introduce one-to-many image translation between synthetic and real images. Evaluation experiments on an appearance-based gaze estimation task demonstrate that the proposed ideas can improve the prediction accuracy and our full method can outperform existing S+U learning methods.

DOI： 10.1109/ICIP40778.2020.9191100

その他リンク： https://dblp.uni-trier.de/db/conf/icip/icip2020.html#MitsuzumiIKN20

First-person Video Analysis for Evaluating Skill Level in the Humanitude Tender-Care Technique

Atsushi Nakazawa, Yu Mitsuzumi, Yuki Watanabe, Ryo Kurazume, Sakiko Yoshikawa, Miwako Honda

Journal of Intelligent and Robotic Systems: Theory and Applications 98 ( 1 ) 103 - 118 2020年4月

掲載種別：研究論文（学術雑誌）

In this paper, we describe a wearable first-person video (FPV) analysis system for evaluating the skill levels of caregivers. This is a part of our project that aims to quantize and analyze the tender-care technique known as Humanitude by using wearable sensing and AI technology devices. Using our system, caregivers can evaluate and elevate their care levels by themselves. From the FPVs of care sessions taken by wearable cameras worn by caregivers, we obtained the 3D facial distance, pose and eye-contact states between caregivers and receivers by using facial landmark detection and deep neural network (DNN)-based eye contact detection. We applied statistical analysis to these features and developed algorithms that provide scores for tender-care skill. In experiments, we first evaluated the performance of our DNN-based eye contact detection by using eye contact datasets prepared from YouTube videos and FPVs that assume conversational scenes. We then performed skill evaluations by using Humanitude training scenes involving three novice caregivers, two Humanitude experts and seven middle-level students. The results showed that our eye contact detection outperformed existing methods and that our skill evaluations can estimate the care skill levels.

DOI： 10.1007/s10846-019-01052-8

Real-time surgical needle detection using region-based convolutional neural networks

Atsushi Nakazawa, Kanako Harada, Mamoru Mitsuishi, Pierre Jannin

International Journal of Computer Assisted Radiology and Surgery 15 ( 1 ) 41 - 47 2020年1月

掲載種別：研究論文（学術雑誌）

Objective: Conventional surgical assistance and skill analysis for suturing mostly focus on the motions of the tools. As the quality of the suturing is determined by needle motions relative to the tissues, having knowledge of the needle motion would be useful for surgical assistance and skill analysis. As the first step toward demonstrating the usefulness of the knowledge of the needle motion, we developed a needle detection algorithm. Methods: Owing to the small needle size, attaching sensors to it is difficult. Therefore, we developed a real-time video-based needle detection algorithm using a region-based convolutional neural network. Results: Our method successfully detected the needle with an average precision of 89.2%. The needle was robustly detected even when the needle was heavily occluded by the tools and/or the blood vessels during microvascular anastomosis. However, there were some incorrect detections, including partial detection. Conclusion: To the best of our knowledge, this is the first time deep neural networks have been applied to real-time needle detection. In the future, we will develop a needle pose estimation algorithm using the predicted needle location toward computer-aided surgical assistance and surgical skill analysis.

DOI： 10.1007/s11548-019-02050-9

その他リンク： https://dblp.uni-trier.de/db/journals/cars/cars15.html#NakazawaHMJ20

Eye Contact Detection from Third Person Video

Yuki Ohshima, Atsushi Nakazawa

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12047 LNCS 667 - 677 2020年

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Springer

Eye contact is fundamental for human communication and social interactions; therefore much effort has been made to develop automated eye-contact detection using image recognition techniques. However, existing methods use first-person-videos (FPV) that need participants to equip wearable cameras. In this work, we develop an novel eye contact detection algorithm taken from normal viewpoint (third person video) assuming the scenes of conversations or social interactions. Our system have high affordability since it does not require special hardware or recording setups, moreover, can use pre-recorded videos such as Youtube and home videos. In designing algorithm, we first develop DNN-based one-sided gaze estimation algorithms which output the states whether the one subject looks at another. Afterwards, eye contact is found at the frame when the pair of one-sided gaze happens. To verify the proposed algorithm, we generate third-person eye contact video dataset using publicly available videos from Youtube. As the result, proposed algorithms performed 0.775 in precision and 0.671 in recall, while the existing method performed 0.484 in precision and 0.061 in recall, respectively.

DOI： 10.1007/978-3-030-41299-9_52

その他リンク： https://dblp.uni-trier.de/db/conf/acpr/acpr2019-2.html#OhshimaN19

傾聴感を高めるためのアンドロイドの瞬目と頷きの模倣動作生成

湯口彰重, 高松淳, 中澤篤志, 小笠原司

ロボティクス・メカトロニクス講演会講演概要集 2020 1P2-E16 2020年

An examination of gaze during conversation for designing culture-based robot behavior

Louisa Hardjasa, Atsushi Nakazawa

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12194 LNCS 475 - 488 2020年

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Springer

Gaze behavior, including eye contact and gaze direction, is an essential component of non-verbal communication, helping to facilitate human-to-human conversation in ways that have often been thought of as universal and innate. However, these have been shown to be influenced partially by cultural norms and background, and despite this, the majority of social robots do not have any cultural-based non-verbal behaviors and several lack any directional gaze capabilities at all. This study aims to observe how different gaze behaviors manifest during conversation as a function of culture as well as exposure to other cultures, by examining differences in behaviors such as duration of direct gaze, duration and direction of averted gaze, and average number of shifts in gaze, with the objective of establishing a baseline of Japanese gaze behavior to be implemented into a social robot. Japanese subjects were found to have much more averted gaze during a task that involves thinking as opposed to a task focused on communication. Subjects with significant experience living overseas were found to have different directional gaze patterns from subjects with little to no overseas experience, implying that non-verbal behavior patterns can change with exposure to other cultures.

DOI： 10.1007/978-3-030-49570-1_33

その他リンク： https://dblp.uni-trier.de/db/conf/hci/hci2020-14.html#HardjasaN20

Comparison of Gaze Skills Between Expert and Novice in Elderly Care

Miyuki Iwamoto, Atsushi Nakazawa

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12208 LNCS 91 - 100 2020年

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Springer

It is known that a person, when ignored by others, causes various negative reactions and enhances aggressive behavior and Self-destructive behavior [1–3]. The same goes for elderly people with dementia, therefore, depending on the care may cause fear and confusion for elderly people with dementia [4]. At the same time, care for the elderly with dementia has a great mental and physical burden on caregivers. For this reason, the turnover rate has increased and it has become difficult to provide adequate care [5–8]. For these problems, as one of the dementia care is gaining attention Humanitude [4, 10]. Humanitude consists of four skills: “see”, “touch”, “speak”, and “stand” [9–11]. So, our research focus on “see”, one of the basic skills. A person’s gaze is generally directed to an object of interest or attention. The gaze is extremely useful information for estimating the mind of another person [3, 12, 13]. Therefore, we let a caregiver wear the first person camera, and four types of “seeing” behavior patterns during the oral care (care receiver → caregiver, care receiver ← caregiver, mutual gaze, none) measured. We compared the differences between Humanitude experts and novice. There was a large difference in the frequency and time of mutual gaze between expert and novice caregiver for care receiver. The act of matching the sight of eyes is an act of not ignoring the other person, indicating an interest in the care receiver, and it is considered that for the care receiver, the anxiety and fear during care are reduced.

DOI： 10.1007/978-3-030-50249-2_7

その他リンク： https://dblp.uni-trier.de/db/conf/hci/hci2020-28.html#IwamotoN20

Evaluating Imitation of Human Eye Contact and Blinking Behavior Using an Android for Human-like Communication

Tetsuya Sano, Akishige Yuguchi, Gustavo Alfonso Garcia Ricardez, Jun Takamatsu, Atsushi Nakazawa, Tsukasa Ogasawara

2019 28th IEEE International Conference on Robot and Human Interactive Communication, RO-MAN 2019 1 - 6 2019年10月

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：IEEE

The appearance of android robots is very similar to that of human beings. From their appearance, we expect that androids might provide us with high-level communication. The imitation of human behavior gives us the feeling of natural behavior even if we do not know what drives high-level communication. In this paper, we evaluate the imitation of human eye behavior by an android. We consider that the android imitates human eye behavior while explaining some research topic and a person acts as a listener. Then, we construct a method to imitate the eye behavior obtained from eye trackers. For the evaluation, we asked seventeen male subjects for their subjective evaluation and compared the imitation with an android that controlled eye-contact duration and eyeblinks by editing the imitation or programming rule-based behavior. From the results, we found out that 1) the rule-based behaviors kept human-likeness, 2) 3-second eye contact obtained better scores regardless of the imitation-based or rule-based eye behavior, and 3) the subjects might regard the longer eyeblinks as voluntary eyeblinks, with the intention to break eye contacts.

DOI： 10.1109/RO-MAN46459.2019.8956387

その他リンク： https://dblp.uni-trier.de/db/conf/ro-man/ro-man2019.html#SanoYRTNO19

First-person camera system to evaluate tender dementia-care skill

Atsushi Nakazawa, Miwako Honda

Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019 4435 - 4442 2019年10月

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：IEEE

In this paper, we describe a wearable first-person video (FPV) analysis system for evaluating the skill levels of tender dementia-care technique. Using this system, caregivers can evaluate and elevate their care levels by themselves using the systems' feedbacks. From the FPVs of care sessions taken by wearable cameras worn by caregivers, we obtained the 3D facial distance, pose and eye-contact states between caregivers and receivers by using facial landmark detection and deep neural network (DNN)-based eye contact detection. We applied statistical analysis to these features and developed algorithms that provide scores for tender-care skill. To find and confirm our idea, we conducted chronological study to observe the progression of tender care-skill learning using care learners. First, we took FPVs while care training scenes involving novice caregivers, tender-care experts and middle-level students, and found major behavioural differences among them. Second, we performed the same experiments for the participants before and after training sessions of the care. As the result, we found the same behavioural difference between 1) novices and experts and 2) novices before and after taking training sessions. These results indicate that our FPV-based behavior analysis can evaluate the skill progression of the tender dementia-care.

DOI： 10.1109/ICCVW.2019.00544

その他リンク： https://dblp.uni-trier.de/db/conf/iccvw/iccvw2019.html#NakazawaH19

HMD-based cover test system for the diagnosis of ocular misalignment

Noriyuki Uchida, Kayoko Takatuka, Hisaaki Yamaba, Atsushi Nakazawa, Masayuki Mukunoki, Naonobu Okazaki

Artificial Life and Robotics 24 ( 3 ) 332 - 337 2019年9月

掲載種別：研究論文（学術雑誌）

The diagnosis of ocular misalignment is difficult and needs examination by ophthalmologists and orthoptists. However, there are not enough qualified personnel to perform such diagnoses. The eye position check is in part systematized. With this check system, we can detect not only the symptoms but also the angle and the extent of strabismus. However, the types of strabismus that can be detected with this technique are limited to exotropia. The purpose of this study is to develop a simplified check system to screen at least the presence of strabismus apart from the type of strabismus or amount of ocular deviation. First, we digitalized the check process. Specifically, we digitized the elemental technology, i.e., the cover–uncover function, required for automation of the typical cover test for eye position check. Furthermore, we developed and implemented an abnormality determination process and evaluated the performance of the system through this experiment, the results of which indicated a higher detection capability than the conventional cover test performed by ophthalmologists and orthoptists.

DOI： 10.1007/s10015-018-0520-4

Towards overhead semantic part segmentation of workers in factory environments from depth images using a FCN

Masakazu Yoshimura, Murilo M. Marinho, Atsushi Nakazawa, Kanako Harada, Mamoru Mitsuishi, Takuya Maekawa, Yasuo Namioka

2019 IEEE International Conference on Cyborg and Bionic Systems, CBS 2019 204 - 209 2019年9月

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：IEEE

In production lines, human workers assemble and/or inspect products in a predetermined process flow. Training new workers is a complex process with varying results. To track the workers motion in a way to better understand human skill, an overhead semantic part segmentation of workers is desired. For this purpose, in this work, we propose a fully-convolutional neural-network model paired with four proposed augmentation strategies. Artificial depth images were used as training data and the augmentation strategies were essential in aiding the network to generalize to the real images. The proposed method was evaluated in two tasks with different backgrounds: a part assembly task and a quality check task. We improved the F-measure by 12% in the part assembly task and 4% in the quality check task when compared to our previous work.

DOI： 10.1109/CBS46900.2019.9114417

その他リンク： https://dblp.uni-trier.de/db/conf/cbs/cbs2019.html#YoshimuraMNHMMN19

Eye Movement on Blank Displays: Effects of Difficulty in Preceding Visual Tasks and Mental State

Ayumi Takemoto, Atsushi Nakazawa, Takatsune Kumada

PERCEPTION 48 220 - 220 2019年9月

Robust Pupil Segmentation and Center Detection from Visible Light Images Using Convolutional Neural Network 査読

Kazunari Kitazumi, Atsushi Nakazawa

Proceedings - 2018 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2018 862 - 868 2019年1月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：IEEE

In this paper, we present a robust pupil detection method from visible light (VL) images using a convolutional neural network (CNN). In contrast to existing pupil detection algorithms, our method does not require infrared (IR) illuminations and cameras, and robustly works even for the images of dark-brown irises (black eyes) and/or images including strong corneal reflections such as outdoor scenes. Thus it has potential application scenarios. Our method first detects an eye region from an input image and applies CNN-based pupil segmentation which has a composition and decomposition structure. To learn the relationship between visible eye images and pupil segments, we construct two datasets and augmentation algorithms. One dataset is based on an existing eye image dataset (UBIRIS.v2), and the other consists of images taken by ourselves by using a corneal imaging camera that can take eye images in mobile environments. Applying color and corneal reflection augmentations to these image datasets and using them for learning the CNN, we built a robust pupil segmentation neural network. The performance is evaluated in three ways. First, we evaluate the segmentation accuracy. Second, we evaluate the pupil center detection accuracy using GI4E facial image sets. Third, we developed an eye gaze tracking (EGT) algorithm that uses the pupil detection and evaluated its accuracy. From the result, the proposed method detects pupil centers more accuracy than the state-of-the arts, and shows similar EGT accuracy to the commercial systems that use IR-active lighting setups.

DOI： 10.1109/SMC.2018.00154

その他リンク： https://dblp.uni-trier.de/db/conf/smc/smc2018.html#KitazumiN18

Eye Contact Detection Algorithms Using Deep Learning and Generative Adversarial Networks 査読

Yu Mitsuzumi, Atsushi Nakazawa

Proceedings - 2018 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2018 3927 - 3931 2019年1月

掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：IEEE

Eye contact (mutual gaze) is a foundation of human communication and social interactions; therefore, it is studied in many fields such as psychology, social science, and medicine. Our group have been studied wearable vision-based eye contact detection techniques using a first person camera for the purpose of evaluating the gaze skills in the tender dementia care. In this work, we search for deep learning-based eye contact detection techniques from small number of labeled images. We implemented and tested two eye contact detection algorithms: naïve deep-learning-based algorithm and generative adversarial networks (GAN)-based semi supervised learning (SSL) algorithm. These methods are learned and verified by using Columbia Gaze Dataset, Facescrub and our original datasets. The results show the effectiveness and limitations of the deep-learning-based and GAN-based approaches. Interestingly, we found the bilateral difference of the accuracy of eye contact detection with respect to the facial pose with respect to the camera, which is expected to be caused by the learning datasets.

DOI： 10.1109/SMC.2018.00666

その他リンク： https://dblp.uni-trier.de/db/conf/smc/smc2018.html#MitsuzumiN18

Spatio-temporal eye contact detection combining CNN and LSTM

Yuki Watanabe, Atsushi Nakazawa, Yu Mitsuzumi, Toyoaki Nishida

PROCEEDINGS OF MVA 2019 16TH INTERNATIONAL CONFERENCE ON MACHINE VISION APPLICATIONS (MVA) 1 - 7 2019年

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：IEEE

Eye contact (mutual gaze) is fundamental for human communication and social interactions; therefore, it is studied in many fields. To support the study of eye contact, much effort has been made to develop automated eye-contact detection using image recognition techniques. In recent years, convolutional neural network (CNN) based eye-contact detection techniques are becoming popular due to their performance; however, they mainly use single frame for recognition. Eye contact is a human communication behavior, so temporal information, such as temporal eye images and facial poses, is important to increase the accuracy of eye-contact detection. We incorporate temporal information into eye-contact detection by using temporal neural network structures that combine CNNs and long short-term memory (LSTM). We tested several network combinations of CNNs and LSTM and found the best solution that uses the outputs of CNNs as well as the cell state vectors of LSTM in the fully connected layers. We prepared two types of eye contact video datasets. One dataset is based on online videos, and the other was taken by a first-person camera in assumed conversational scenarios. The results show that our method is better than the approaches that use single frames. Namely, our method performs 0.8781, while the existing method (DeepEC) performed 0.8319, in F-1-score.

DOI： 10.23919/MVA.2019.8757989

その他リンク： https://dblp.uni-trier.de/db/conf/mva/mva2019.html#WatanabeNMN19

General Improvement Method of Specular Component Separation Using High-Emphasis Filter and Similarity Function

Takahisa Yamamoto, Atsushi Nakazawa

ITE TRANSACTIONS ON MEDIA TECHNOLOGY AND APPLICATIONS 7 ( 2 ) 92 - 102 2019年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：INST IMAGE INFORMATION & TELEVISION ENGINEERS

Separating reflection components is a fundamental problem in computer vision and useful for many applications such as image quality. We propose a novel method that improves the accuracy of separating reflection components from a single image. Although several algorithms for separating reflection components have been proposed, our method can additionally improve the accuracy based on their results. First, we obtain diffuse and specular components by using an existing algorithm. Then, we apply a high-emphasis filter for each component. Since the responses of the high-emphasis filter where the separation fails become larger than the original values, we can detect erroneous pixels. Thus, we replace separation results of these erroneous pixels with results of other reference pixels from the image considering the similarity between the target and reference pixels. Experimental results show that our method can improve at most 13.61 dB in terms of the Peak Signal-to-Noise Ratio (PSNR).

DOI： 10.3169/mta.7.92

　前のページ - 次のページ　

1
2
3
4
5
6