• ETRI Journal
• /
• v.46 no.1
• /
• pp.118-126
• /
• 2024

In human conversations, listeners often utilize brief backchannels such as "uh-huh" or "yeah." Timely backchannels are crucial to understanding and increasing trust among conversational partners. In human-machine conversation systems, users can engage in natural conversations when a conversational agent generates backchannels like a human listener. We propose a method that simultaneously predicts backchannels and recognizes speech in real time. We use a streaming transformer and adopt multitask learning for concurrent backchannel prediction and speech recognition. The experimental results demonstrate the superior performance of our method compared with previous works while maintaining a similar single-task speech recognition performance. Owing to the extremely imbalanced training data distribution, the single-task backchannel prediction model fails to predict any of the backchannel categories, and the proposed multitask approach substantially enhances the backchannel prediction performance. Notably, in the streaming prediction scenario, the performance of backchannel prediction improves by up to 18.7% compared with existing methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.2
• /
• pp.751-770
• /
• 2019

Action recognition is an essential task in computer vision due to the variety of prospective applications, such as security surveillance, machine learning, and human-computer interaction. The availability of more video data than ever before and the lofty performance of deep convolutional neural networks also make it essential for action recognition in video. Unfortunately, limited crafted video features and the scarcity of benchmark datasets make it challenging to address the multi-person action recognition task in video data. In this work, we propose a deep convolutional neural network-based Effective Hybrid Learning (EHL) framework for two-person interaction classification in video data. Our approach exploits a pre-trained network model (the VGG16 from the University of Oxford Visual Geometry Group) and extends the Faster R-CNN (region-based convolutional neural network a state-of-the-art detector for image classification). We broaden a semi-supervised learning method combined with an active learning method to improve overall performance. Numerous types of two-person interactions exist in the real world, which makes this a challenging task. In our experiment, we consider a limited number of actions, such as hugging, fighting, linking arms, talking, and kidnapping in two environment such simple and complex. We show that our trained model with an active semi-supervised learning architecture gradually improves the performance. In a simple environment using an Intelligent Technology Laboratory (ITLab) dataset from Inha University, performance increased to 95.6% accuracy, and in a complex environment, performance reached 81% accuracy. Our method reduces data-labeling time, compared to supervised learning methods, for the ITLab dataset. We also conduct extensive experiment on Human Action Recognition benchmarks such as UT-Interaction dataset, HMDB51 dataset and obtain better performance than state-of-the-art approaches.

• Animal cells and systems
• /
• v.12 no.1
• /
• pp.11-14
• /
• 2008

Finding a way to strengthen human cognitive functions, such as learning and memory, has been of great concern since the moment people realized that these functions can be affected and even altered by certain chemicals. Since then, plenty of endeavors have been made to look for safe ways of improving cognitive performances without adverse side-effects. Unfortunately, most of these efforts have turned out to be unsuccessful until now. In this study, we examine the effect of a natural compound, idesolide, on hippocampus-dependent recognition memory. We demonstrate that idesolide is effective in the enhancement of recognition memory, as measured by a novel object recognition task. Thus, idesolide might serve as a novel therapeutic medication for the treatment of memoryrelated brain anomalies such as mild cognitive impairment(MCI) and Alzheimer's disease.

• Journal of the Korean Society for Precision Engineering
• /
• v.20 no.1
• /
• pp.99-106
• /
• 2003

Uninterrupted power supply has become indispensable during the maintenance task of active electric power lines as a result of today's highly information-oriented society and increasing demand of electric utilities. The maintenance task has the risk of electric shock and the danger of falling from high place. Therefore it is necessary to realize an autonomous robot system. In order to realize these tasks autonomously, the there dimensional position of target object such as electric line and the stand of insulator must be recognized accurately and rapidly. The insertion task of an electric line into a sleeve is selected as the typical task of the maintenance of active electric power distribution lines in this paper. A modified hough transform is applied to the recognition of the center of electric line and optimal target position calculation method is newly derived in order to recognize the center 3 dimensional position of the electric line. By the proposed method, it is proved that the center position of the electric line can be recognized without respect to the noise of image and the shape of electric lines and the insertion task of an electric tine is realized.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.2
• /
• pp.483-503
• /
• 2014

In this paper, human action recognition using pyramid histograms of oriented gradients and collaborative multi-task learning is proposed. First, we accumulate global activities and construct motion history image (MHI) for both RGB and depth channels respectively to encode the dynamics of one action in different modalities, and then different action descriptors are extracted from depth and RGB MHI to represent global textual and structural characteristics of these actions. Specially, average value in hierarchical block, GIST and pyramid histograms of oriented gradients descriptors are employed to represent human motion. To demonstrate the superiority of the proposed method, we evaluate them by KNN, SVM with linear and RBF kernels, SRC and CRC models on DHA dataset, the well-known dataset for human action recognition. Large scale experimental results show our descriptors are robust, stable and efficient, and outperform the state-of-the-art methods. In addition, we investigate the performance of our descriptors further by combining these descriptors on DHA dataset, and observe that the performances of combined descriptors are much better than just using only sole descriptor. With multimodal features, we also propose a collaborative multi-task learning method for model learning and inference based on transfer learning theory. The main contributions lie in four aspects: 1) the proposed encoding the scheme can filter the stationary part of human body and reduce noise interference; 2) different kind of features and models are assessed, and the neighbor gradients information and pyramid layers are very helpful for representing these actions; 3) The proposed model can fuse the features from different modalities regardless of the sensor types, the ranges of the value, and the dimensions of different features; 4) The latent common knowledge among different modalities can be discovered by transfer learning to boost the performance.

• Journal of Digital Convergence
• /
• v.17 no.12
• /
• pp.243-248
• /
• 2019

Multi-Task Learning(MTL) is a training method that trains a single neural network with multiple tasks influences each other. In this paper, we compare performance of MTL Named entity recognition(NER) model trained with Korean traditional culture corpus and other NER model. In training process, each Bi-LSTM layer of Part of speech tagging(POS-tagging) and NER are propagated from a Bi-LSTM layer to obtain the joint loss. As a result, the MTL based Bi-LSTM model shows 1.1%~4.6% performance improvement compared to single Bi-LSTM models.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.5
• /
• pp.388-402
• /
• 2003

In this paper, we carried out the word, 4 continuous digits. continuous, and task-independent word recognition experiments to verify the effectiveness of the re-defined phoneme-likely units (PLUs) for the phonetic decision tree based HM-Net (Hidden Markov Network) context-dependent (CD) acoustic modeling in Korean appropriately. In case of the 48 PLUs, the phonemes /ㅂ/, /ㄷ/, /ㄱ/ are separated by initial sound, medial vowel, final consonant, and the consonants /ㄹ/, /ㅈ/, /ㅎ/ are also separated by initial sound, final consonant according to the position of syllable, word, and sentence, respectively. In this paper. therefore, we re-define the 39 PLUs by unifying the one phoneme in the separated initial sound, medial vowel, and final consonant of the 48 PLUs to construct the CD acoustic models effectively. Through the experimental results using the re-defined 39 PLUs, in word recognition experiments with the context-independent (CI) acoustic models, the 48 PLUs has an average of 7.06%, higher recognition accuracy than the 39 PLUs used. But in the speaker-independent word recognition experiments with the CD acoustic models, the 39 PLUs has an average of 0.61% better recognition accuracy than the 48 PLUs used. In the 4 continuous digits recognition experiments with the liaison phenomena. the 39 PLUs has also an average of 6.55% higher recognition accuracy. And then, in continuous speech recognition experiments, the 39 PLUs has an average of 15.08% better recognition accuracy than the 48 PLUs used too. Finally, though the 48, 39 PLUs have the lower recognition accuracy, the 39 PLUs has an average of 1.17% higher recognition characteristic than the 48 PLUs used in the task-independent word recognition experiments according to the unknown contextual factor. Through the above experiments, we verified the effectiveness of the re-defined 39 PLUs compared to the 48PLUs to construct the CD acoustic models in this paper.

• Korean Journal of Biological Psychiatry
• /
• v.18 no.4
• /
• pp.239-244
• /
• 2011

Objectives Schizophrenic patients have been shown to be impaired in both emotional self-awareness and recognition of others' facial emotions. Alexithymia refers to the deficits in emotional self-awareness. The relationship between alexithymia and recognition of others' facial emotions needs to be explored to better understand the characteristics of emotional deficits in schizophrenic patients. Methods Thirty control subjects and 31 schizophrenic patients completed the Toronto Alexithymia Scale-20-Korean version (TAS-20K) and facial emotion recognition task. The stimuli in facial emotion recognition task consist of 6 emotions (happiness, sadness, anger, fear, disgust, and neutral). Recognition accuracy was calculated within each emotion category. Correlations between TAS-20K and recognition accuracy were analyzed. Results The schizophrenic patients showed higher TAS-20K scores and lower recognition accuracy compared with the control subjects. The schizophrenic patients did not demonstrate any significant correlations between TAS-20K and recognition accuracy, unlike the control subjects. Conclusions The data suggest that, although schizophrenia may impair both emotional self-awareness and recognition of others' facial emotions, the degrees of deficit can be different between emotional self-awareness and recognition of others' facial emotions. This indicates that the emotional deficits in schizophrenia may assume more complex features.

• MALSORI
• /
• no.60
• /
• pp.29-45
• /
• 2006

We investigated whether the first syllable plays an important role in lexical access in Korean visual word recognition. To do so, one lexical decision task (LDT) and two form primed LDT experiments examined the nature of the syllabic neighborhood effect. In Experiment 1, the syllabic neighborhood density and the syllabic neighborhood frequency was manipulated. The results showed that lexical decision latencies were only influenced by the syllabic neighborhood frequency. The purpose of experiment 2 was to confirm the results of experiment 1 with form-primed LDT task. The lexical decision latency was slower in form-related condition compared to form-unrelated condition. The effect of syllabic neighborhood density was significant only in form-related condition. This means that the first syllable plays an important role in the sub-lexical process. In Experiment 3, we conducted another form-primed LDT task manipulating the number of syllabic neighbors in words with higher frequency neighborhood. The interaction of syllabic neighborhood density and form relation was significant. This result confirmed that the words with higher frequency neighborhood are more inhibited by neighbors sharing the first syllable than words with no higher frequency neighborhood in the lexical level. These findings suggest that the first syllable is the unit of neighborhood and the unit of representation in sub-lexical representation is syllable in Korea.

• Journal of Information Processing Systems
• /
• v.17 no.2
• /
• pp.337-351
• /
• 2021

Automatically recognizing facial expressions in video sequences is a challenging task because there is little direct correlation between facial features and subjective emotions in video. To overcome the problem, a video facial expression recognition method using spatiotemporal recurrent neural network and feature fusion is proposed. Firstly, the video is preprocessed. Then, the double-layer cascade structure is used to detect a face in a video image. In addition, two deep convolutional neural networks are used to extract the time-domain and airspace facial features in the video. The spatial convolutional neural network is used to extract the spatial information features from each frame of the static expression images in the video. The temporal convolutional neural network is used to extract the dynamic information features from the optical flow information from multiple frames of expression images in the video. A multiplication fusion is performed with the spatiotemporal features learned by the two deep convolutional neural networks. Finally, the fused features are input to the support vector machine to realize the facial expression classification task. The experimental results on cNTERFACE, RML, and AFEW6.0 datasets show that the recognition rates obtained by the proposed method are as high as 88.67%, 70.32%, and 63.84%, respectively. Comparative experiments show that the proposed method obtains higher recognition accuracy than other recently reported methods.