• Journal of the Korea Society of Computer and Information
• /
• v.28 no.9
• /
• pp.47-54
• /
• 2023

Infants utilize crying as a non-verbal means of communication [1]. However, deciphering infant cries presents challenges. Extensive research has been conducted to interpret infant cry audios [2,3]. This paper proposes the classification of infant cries using 3D feature vectors augmented with various audio data techniques. A total of 5 classes (belly pain, burping, discomfort, hungry, tired) are employed in the study dataset. The data is augmented using 5 techniques (Pitch, Tempo, Shift, Mixup-noise, CutMix). Tempo, Shift, and CutMix augmentation techniques demonstrated improved performance. Ultimately, applying effective data augmentation techniques simultaneously resulted in a 17.75% performance enhancement compared to models using single feature vectors and original data.

• Journal of Biomedical Engineering Research
• /
• v.31 no.6
• /
• pp.481-486
• /
• 2010

In this paper, we suggest an advanced algorithm, to recognize pedestrian/non-pedestrian using differential haar-like feature, which applies Adaboost algorithm to make a strong classification from weak classifications. First, we extract two feature vectors: horizontal haar-like feature and vertical haar-like feature. For the next, we calculate the proposed feature vector using differential haar-like method. And then, a strong classification needs to be obtained from weak classifications for composite recognition method using the differential area of horizontal and vertical haar-like. In the proposed method, we use one feature vector and one strong classification for the first stage of recognition. Based on our experiment, the proposed algorithm shows higher recognition rate compared to the traditional method for the pedestrian and non-pedestrian.

• ETRI Journal
• /
• v.41 no.2
• /
• pp.235-241
• /
• 2019

In this paper, we propose a rank-weighted reconstruction feature to improve the robustness of a feed-forward deep neural network (FFDNN)-based acoustic model. In the FFDNN-based acoustic model, an input feature is constructed by vectorizing a submatrix that is created by slicing the feature vectors of frames within a context window. In this type of feature construction, the appropriate context window size is important because it determines the amount of trivial or discriminative information, such as redundancy, or temporal context of the input features. However, we ascertained whether a single parameter is sufficiently able to control the quantity of information. Therefore, we investigated the input feature construction from the perspectives of rank and nullity, and proposed a rank-weighted reconstruction feature herein, that allows for the retention of speech information components and the reduction in trivial components. The proposed method was evaluated in the TIMIT phone recognition and Wall Street Journal (WSJ) domains. The proposed method reduced the phone error rate of the TIMIT domain from 18.4% to 18.0%, and the word error rate of the WSJ domain from 4.70% to 4.43%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.3
• /
• pp.894-907
• /
• 2022

The features extracted by convolutional neural networks are more descriptive of images than traditional features, and their convolutional layers are more suitable for retrieving images than are fully connected layers. The convolutional layer features will consume considerable time and memory if used directly to match an image. Therefore, this paper proposes a feature weighting and region integration method for convolutional layer features to form global feature vectors and subsequently use them for image matching. First, the 3D feature of the last convolutional layer is extracted, and the convolutional feature is subsequently weighted again to highlight the edge information and position information of the image. Next, we integrate several regional eigenvectors that are processed by sliding windows into a global eigenvector. Finally, the initial ranking of the retrieval is obtained by measuring the similarity of the query image and the test image using the cosine distance, and the final mean Average Precision (mAP) is obtained by using the extended query method for rearrangement. We conduct experiments using the Oxford5k and Paris6k datasets and their extended datasets, Paris106k and Oxford105k. These experimental results indicate that the global feature extracted by the new method can better describe an image.

• Journal of Internet Computing and Services
• /
• v.8 no.3
• /
• pp.19-32
• /
• 2007

This paper present a new face recognition method based on LVQ neural net to construct a real time face recognition system. The previous researches which used PCA, LDA combined neural net usually need much time in training neural net. The supervised LVQ neural net needs much less time in training and can maximize the separability between the classes. In this paper, the proposed method transforms the input face image by PCA and LDA sequentially into low-dimension feature vectors and recognizes the face through LVQ neural net. In order to make the system robust to external light variation, light compensation is performed on the detected face by max-min normalization method as preprocessing. PCA and LDA transformations are applied to the normalized face image to produce low-level feature vectors of the image. In order to determine the initial centers of LVQ and speed up the convergency of the LVQ neural net, the K-Means clustering algorithm is adopted. Subsequently, the class representative vectors can be produced by LVQ2 training using initial center vectors. The face recognition is achieved by using the euclidean distance measure between the center vector of classes and the feature vector of input image. From the experiments, we can prove that the proposed method is more effective in the recognition ratio for the cases of still images from ORL database and sequential images rather than using conventional PCA of a hybrid method with PCA and LDA.

• Journal of The Korean Astronomical Society
• /
• v.42 no.3
• /
• pp.61-69
• /
• 2009

Transverse velocity vectors can be determined from a pair of images successively taken with a time interval using an optical flow technique. We have tested the performance of the new technique called NAVE (non-linear affine velocity estimator) recently implemented by Chae & Sakurai using real image data taken by the Narrowband Filter Imager (NFI) of the Solar Optical Telescope (SOT) aboard the Hinode satellite. We have developed two methods of estimating the errors in the determination of velocity vectors, one resulting from the non-linear fitting ${\sigma}_{\upsilon}$ and the other ${\epsilon}_u$ resulting from the statistics of the determined velocity vectors. The real error is expected to be somewhere between ${\sigma}_{\upsilon}$ and ${\epsilon}_u$. We have investigated the dependence of the determined velocity vectors and their errors on the different parameters such as the critical speed for the subsonic filtering, the width of the localizing window, the time interval between two successive images, and the signal-to-noise ratio of the feature. With the choice of $v_{crit}$ = 2 pixel/step for the subsonic filtering, and the window FWHM of 16 pixels, and the time interval of one step (2 minutes), we find that the errors of velocity vectors determined using the NAVE range from around 0.04 pixel/step in high signal-to-noise ratio features (S/N $\sim$ 10), to 0.1 pixel/step in low signa-to-noise ratio features (S/N $\sim$ 3) with the mean of about 0.06 pixel/step where 1 pixel/step corresponds roughly to 1 km/s in our case.

• Journal of IKEEE
• /
• v.13 no.4
• /
• pp.43-46
• /
• 2009

This paper deals with feeling recognition from people's voice to fine feature vectors. Voice signals include the people's own information and but also people's feelings and fatigues. So, many researches are being progressed to fine the feelings from people's voice. In this paper, We analysis Selectable Mode Vocoder(SMV) that is one of the standard 3GPP2 codecs of ETSI. From the analyzed result, we propose voices features for recognizing feelings. And then, feeling recognition algorithm based on gaussian mixture model(GMM) is proposed. It uses feature vectors is suggested. We verify the performance of this algorithm from changing the mixture component.

• Journal of Korea Multimedia Society
• /
• v.3 no.3
• /
• pp.241-252
• /
• 2000

In this paper, a translation, rotation and scale invariant system for the recognition of closed 2-D images using the bispectrum of a contour sequence and a weighted fuzzy classifier is derived and compared with the recognition process using one of the competitive neural algorithm, called a LVQ( Loaming Vector Quantization). The bispectrum based on third order cumulants is applied to the contour sequences of an image to extract fifteen feature vectors for each planar image. These bispectral feature vectors, which are invariant to shape translation, rotation and scale transformation, can be used to the represent two-dimensional planar images and are fed into a weighted fuzzy classifier. The experimental processes with eight different shapes of aircraft images are presented to illustrate a relatively high performance of the proposed recognition system.

• Science of Emotion and Sensibility
• /
• v.22 no.3
• /
• pp.55-64
• /
• 2019

Stroke is a health problem experienced by many elderly people around the world. Stroke has a devastating effect on quality of life, causing death or disability. Hemiplegia is clearly an early sign of a stroke and can be detected through patterns of body balance and gait. The goal of this study was to determine various feature vectors of foot pressure and gait parameters of patients with stroke through the use of a wearable sensor and to compare the gait parameters with those of healthy elderly people. To monitor the participants at all times, we used a simple measuring device rather than a medical device. We measured gait data of 220 healthy people older than 65 years of age and of 63 elderly patients who had experienced stroke less than 6 months earlier. The center of pressure and the acceleration during standing and gait-related tasks were recorded by a wearable insole sensor worn by the participants. Both the average acceleration and the maximum acceleration were significantly higher in the healthy participants (p < .01) than in the patients with stroke. Thus gait parameters are helpful for determining whether they are patients with stroke or normal elderly people.

• The Journal of the Korea Contents Association
• /
• v.9 no.8
• /
• pp.121-128
• /
• 2009

Iris recognition is a biometric technology which can identify a person using the iris pattern. Recently, iris information is used in many fields such as access control and information security. It is important for the iris recognition system to extract the feature which is invariant to changes in iris patterns. Those changes can be occurred by the influence of lights, changes in the size of the pupil. In this paper, we propose a novel method based on vector summation of gradient orientation vectors. Experimental results show that the proposed method reduces processing time with simple vector calculation, requires small feature space and has comparable performance to the well-known previous methods.