• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06b
• /
• pp.231-234
• /
• 2011

Audiovisual-based human emotion recognition can be considered a good approach for multimodal humancomputer interaction. However, the optimal multimodal information fusion remains challenges. In order to overcome the limitations and bring robustness to the interface, we propose a framework of automatic human emotion recognition system from speech and face display. In this paper, we develop a new approach for fusing information in model-level based on the relationship between speech and face expression to detect automatic temporal segments and perform multimodal information fusion.

• Journal of Biomedical Engineering Research
• /
• v.31 no.2
• /
• pp.162-169
• /
• 2010

This paper deals with the recognition algorithm of walking will based on torque estimation. Recently, concern about walking assistant aids is increasing according to the increase in population of elder and handicapped person. However, most of walking aids don't have any actuators for its movement. So, general walking aids have weakness for its movement to upward/download direction of slope. To overcome the weakness of the general walking aids, many researches for active type walking aids are being progressed. Unfortunately it is difficult to control aids during its movement, because it is not easy to recognize user's walking will. Many kinds of methods are proposed to recognize of user's walking will. In this paper, we propose walking will recognition algorithm by using torque estimation from wheels. First, we measure wheel velocity and voltage at the walking aids. From these data, external forces are extracted. And then walking will that is included by walking velocity and direction is estimated. Here, all the processes are verified by simulation and experiment in the real world.

• Proceedings of the Korean Statistical Society Conference
• /
• 2003.10a
• /
• pp.69-72
• /
• 2003

The paper considers partial least squares (PLS) as a new dimension reduction technique for the feature vector to overcome the small sample size problem in face recognition. Principal component analysis (PCA), a conventional dimension reduction method, selects the components with maximum variability, irrespective of the class information. So PCA does not necessarily extract features that are important for the discrimination of classes. PLS, on the other hand, constructs the components so that the correlation between the class variable and themselves is maximized. Therefore PLS components are more predictive than PCA components in classification. The experimental results on Manchester and ORL databases show that PLS is to be preferred over PCA when classification is the goal and dimension reduction is needed.

• MALSORI
• /
• no.53
• /
• pp.103-118
• /
• 2005

The performance of speech recognition system is degraded by mismatch between training and test environments. Recently, Stereo-based Piecewise LInear Compensation for Environments (SPLICE) was introduced to overcome environmental mismatch using stereo data. In this paper, we propose several methods to improve the conventional SPLICE and evaluate them in the Aurora2 task. We generalize SPLICE to compensate for covariance matrix as well as mean vector in the feature space, and thereby yielding the error rate reduction of 48.93%. We also employ the weighted sum of correction vectors using posterior probabilities of all Gaussians, and the error rate reduction of 48.62% is achieved. With the combination of the above two methods, the error rate is reduced by 49.61% from the Aurora2 baseline system.

• Current Optics and Photonics
• /
• v.4 no.5
• /
• pp.421-427
• /
• 2020

A novel privacy-protection device to prevent visual hacking is realized by using a directional backlight and facial recognition. The proposed method is able to overcome the limitations of previous privacy-protection methods that simply restrict the viewing angle to a narrow range. The accuracy of user tracking is accomplished by the combination of a time-of-flight sensor and facial recognition with no restriction of detection range. In addition, an experimental demonstration is provided to verify the proposed scheme.

• Journal of Advanced Information Technology and Convergence
• /
• v.8 no.2
• /
• pp.79-87
• /
• 2018

The current PC user authentication uses character password based on user's knowledge. However, this can easily be exploited by password cracking or key-logging programs. In addition, the use of a difficult password and the periodic change of the password make it easy for the user to mistake exposing the password around the PC because it is difficult for the user to remember the password. In order to overcome this, we propose user gesture recognition and challenge-response authentication. We apply user's hand gesture instead of character password. In the challenge-response method, authentication is performed in the form of responding to a quiz, rather than using the same password every time. To apply the hand gesture to challenge-response authentication, the gesture is recognized and symbolized to be used in the quiz response. So we show that this method can be applied to PC user authentication.

• Journal of Sensor Science and Technology
• /
• v.31 no.1
• /
• pp.16-23
• /
• 2022

Cardiac disease is the most common cause of death worldwide. Therefore, detection and classification of electrocardiogram (ECG) signals are crucial to extend life expectancy. In this study, we aimed to implement an artificial intelligence signal recognition system in field programmable gate array (FPGA), which can recognize patterns of bio-signals such as ECG in edge devices that require batteries. Despite the increment in classification accuracy, deep learning models require exorbitant computational resources and power, which makes the mapping of deep neural networks slow and implementation on wearable devices challenging. To overcome these limitations, spiking neural networks (SNNs) have been applied. SNNs are biologically inspired, event-driven neural networks that compute and transfer information using discrete spikes, which require fewer operations and less complex hardware resources. Thus, they are more energy-efficient compared to other artificial neural networks algorithms.

• International conference on construction engineering and project management
• /
• 2009.05a
• /
• pp.257-262
• /
• 2009

Image processing has been utilized for assessment of infrastructure surface coating conditions for years. However, there is no robust method to overcome the non-uniform illumination problem to date. Therefore, this paper aims to deal with non-uniform illumination problems for bridge coating assessment and to achieve automated rust intensity recognition. This paper starts with selection of the best color configuration for non-uniformly illuminated rust image segmentation. The adaptive-network-based fuzzy inference system (ANFIS) is adopted as the framework to develop the new model, the box-and-ellipse-based neuro-fuzzy approach (BENFA). Finally, the performance of BENFA is compared to the Fuzzy C-Means (FCM) method, which is often used in image recognition, to show the advantage and robustness of BENFA.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.3
• /
• pp.1614-1632
• /
• 2017

Bag of visual words is a popular model in human action recognition, but usually suffers from loss of spatial and temporal configuration information of local features, and large quantization error in its feature coding procedure. In this paper, to overcome the two deficiencies, we combine sparse coding with spatio-temporal pyramid for human action recognition, and regard this method as the baseline. More importantly, which is also the focus of this paper, we find that there is a hierarchical structure in feature vector constructed by the baseline method. To exploit the hierarchical structure information for better recognition accuracy, we propose a tree regularized classifier to convey the hierarchical structure information. The main contributions of this paper can be summarized as: first, we introduce a tree regularized classifier to encode the hierarchical structure information in feature vector for human action recognition. Second, we present an optimization algorithm to learn the parameters of the proposed classifier. Third, the performance of the proposed classifier is evaluated on YouTube, Hollywood2, and UCF50 datasets, the experimental results show that the proposed tree regularized classifier obtains better performance than SVM and other popular classifiers, and achieves promising results on the three datasets.

• Journal of Korea Multimedia Society
• /
• v.8 no.7
• /
• pp.898-905
• /
• 2005

Iris recognition Is that identifies a user based on the unique iris muscle patterns which has the functionalities of dilating or contracting pupil region. Because it is reported that iris recognition is more accurate than other biometries such as face, fingerprint, vein and speaker recognition, iris recognition is widely used in the high security application domain. However, if unnecessary information such as eyelid and eyelash is included in iris region, the error for iris recognition is increased, consequently. In detail, if iris region is used to generate iris code including eyelash and eyelid, the iris codes are also changed and the error rate is increased. To overcome such problem, we propose the method of detecting eyelid by using pyramid searching parabolic deformable template. In addition, we detect the eyelash by using the eyelash mask. Experimental results show that EER(Equal Error Rate) for iris recognition using the proposed algorithm is lessened as much as $0.3\%$ compared to that not using it.