• Journal of KIISE:Software and Applications
• /
• v.31 no.5
• /
• pp.617-624
• /
• 2004

This paper describes a new facial feature localization method that uses Adjacent Depth Differences(ADD) in 3D facial surface. In general, human recognize the extent of deepness or shallowness of region relatively, in depth, by comparing the neighboring depth information among regions of an object. The larger the depth difference between regions shows, the easier one can recognize each region. Using this principal, facial feature extraction will be easier, more reliable and speedy. 3D range images are used as input images. And ADD are obtained by differencing two range values, which are separated at a distance coordinate, both in horizontal and vertical directions. ADD and input image are analyzed to extract facial features, then localized a nose region, which is the most prominent feature in 3D facial surface, effectively and accurately.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.5-7
• /
• 1996

We propose a triangulation method based on stereo vision angles. We setup stereo vision systems which extract the depth information to a moving object by detecting a moving object using difference image method and obtaining the depth information by the triangulation method based on stereo vision angles. The feature point of a moving object is used the geometrical center of the moving object, and the proposed vision system has the accuracy of 0.2mm in the range of 400mm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.8
• /
• pp.2068-2082
• /
• 2023

With the rapid development of deep learning, Depth Map Super-Resolution (DMSR) method has achieved more advanced performances. However, when the upsampling rate is very large, it is difficult to capture the structural consistency between color features and depth features by these DMSR methods. Therefore, we propose a color-image guided DMSR method based on iterative depth feature enhancement. Considering the feature difference between high-quality color features and low-quality depth features, we propose to decompose the depth features into High-Frequency (HF) and Low-Frequency (LF) components. Due to structural homogeneity of depth HF components and HF color features, only HF color features are used to enhance the depth HF features without using the LF color features. Before the HF and LF depth feature decomposition, the LF component of the previous depth decomposition and the updated HF component are combined together. After decomposing and reorganizing recursively-updated features, we combine all the depth LF features with the final updated depth HF features to obtain the enhanced-depth features. Next, the enhanced-depth features are input into the multistage depth map fusion reconstruction block, in which the cross enhancement module is introduced into the reconstruction block to fully mine the spatial correlation of depth map by interleaving various features between different convolution groups. Experimental results can show that the two objective assessments of root mean square error and mean absolute deviation of the proposed method are superior to those of many latest DMSR methods.

• The Journal of the Korea Contents Association
• /
• v.18 no.11
• /
• pp.447-454
• /
• 2018

This paper proposes a depth image generation algorithm of stereo images using a deep learning model composed of a CNN (convolutional neural network). The proposed algorithm consists of a feature extraction unit which extracts the main features of each parallax image and a depth learning unit which learns the parallax information using extracted features. First, the feature extraction unit extracts a feature map for each parallax image through the Xception module and the ASPP(Atrous spatial pyramid pooling) module, which are composed of 2D CNN layers. Then, the feature map for each parallax is accumulated in 3D form according to the time difference and the depth image is estimated after passing through the depth learning unit for learning the depth estimation weight through 3D CNN. The proposed algorithm estimates the depth of object region more accurately than other algorithms.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.256-257
• /
• 2022

In this paper, we introduce a method of extracting three-dimensional feature points through the movement of a single mobile device. Using a monocular camera, a 2D image is acquired according to the camera movement and a baseline is estimated. Perform stereo matching based on feature points. A feature point and a descriptor are acquired, and the feature point is matched. Using the matched feature points, the disparity is calculated and a depth value is generated. The 3D feature point is updated according to the camera movement. Finally, the feature point is reset at the time of scene change by using scene change detection. Through the above process, an average of 73.5% of additional storage space can be secured in the key point database. By applying the algorithm proposed to the depth ground truth value of the TUM Dataset and the RGB image, it was confirmed that the\re was an average distance difference of 26.88mm compared with the 3D feature point result.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.3
• /
• pp.25-35
• /
• 2014

This paper proposes a method of boosting-based classification for the purpose of real-time face detection. The proposed method uses depth images to ensure strong performance of face detection in response to changes in lighting and face size, and uses the depth difference feature to conduct learning and recognition through the I-MCTBoost classifier. I-MCTBoost performs recognition by connecting the strong classifiers that are constituted from weak classifiers. The learning process for the weak classifiers is as follows: first, depth difference features are generated, and eight of these features are combined to form the weak classifier, and each feature is expressed as a binary bit. Strong classifiers undergo learning through the process of repeatedly selecting a specified number of weak classifiers, and become capable of strong classification through a learning process in which the weight of the learning samples are renewed and learning data is added. This paper explains depth difference features and proposes a learning method for the weak classifiers and strong classifiers of I-MCTBoost. Lastly, the paper presents comparisons of the proposed classifiers and the classifiers using conventional MCT through qualitative and quantitative analyses to establish the feasibility and efficiency of the proposed classifiers.

• Journal of Korean Society on Water Environment
• /
• v.35 no.5
• /
• pp.432-438
• /
• 2019

The purpose of this study is to predict Trophic Diatom Index (TDI) in tributaries of the Han River watershed using the random forest algorithm. The one year (2017) and supplied aquatic ecology health data were used. The data includes water quality(BOD, T-N, $NH_3-N$, T-P, $PO_4-P$, water temperature, DO, pH, conductivity, turbidity), hydraulic factors(water width, average water depth, average velocity of water), and TDI score. Seven factors including water temperature, BOD, T-N, $NH_3-N$, T-P, $PO_4-P$, and average water depth are selected by the Correlation Feature Selection. A TDI prediction model was generated by random forest using the seven factors. To evaluate this model, 2017 data set was used first. As a result of the evaluation, $R^2$, % Difference, NSE(Nash-Sutcliffe Efficiency), RMSE(Root Mean Square Error) and accuracy rate show that this model is compatible with predicting TDI. To be more concrete, $R^2$ is 0.93, % Difference is -0.37, NSE is 0.89, RMSE is 8.22 and accuracy rate is 70.4%. Also, additional evaluation using data set more than 17 times the measured point was performed. The results were similar when the 2017 data set were used. The Wilcoxon Signed Ranks Test shows there was no statistically significant difference between actual and predicted data for the 2017 data set. These results can specify the elements which probably affect aquatic ecology health. Also, these will provide direction relative to water quality management for a watershed that must be continuously preserved.

• Journal of Information Processing Systems
• /
• v.11 no.4
• /
• pp.643-654
• /
• 2015

Face recognition under controlled settings, such as limited viewpoint and illumination change, can achieve good performance nowadays. However, real world application for face recognition is still challenging. In this paper, we propose using the combination of Affine Scale Invariant Feature Transform (SIFT) and Probabilistic Similarity for face recognition under a large viewpoint change. Affine SIFT is an extension of SIFT algorithm to detect affine invariant local descriptors. Affine SIFT generates a series of different viewpoints using affine transformation. In this way, it allows for a viewpoint difference between the gallery face and probe face. However, the human face is not planar as it contains significant 3D depth. Affine SIFT does not work well for significant change in pose. To complement this, we combined it with probabilistic similarity, which gets the log likelihood between the probe and gallery face based on sum of squared difference (SSD) distribution in an offline learning process. Our experiment results show that our framework achieves impressive better recognition accuracy than other algorithms compared on the FERET database.

• Journal of Periodontal and Implant Science
• /
• v.35 no.3
• /
• pp.761-776
• /
• 2005

The purpose of the present study was to examine the relationship between the form of the clinical crowns in the maxillary anterior segment and the clinical feature of gingiva such as morphological characteristics and the gingival thickness. Fifty periodontally healthy subjects were clinically examined regarding the probing depth, the thickness of the free gingiva, and the width of the keratinized gingiva. From study models of the maxillary anterior region, the width at cervical third(CW) and the length(CL) of the clinical crown, the papillary height, and the gingival angle of the 6 anterior teeth were measured. Each tooth was classified into 4 groups (longnarrow, NL; narrow, N; wide, W; short-wide, WS) according to CW/CL ratio and all the data were compared between groups NL and WS using independent t-test. Stepwise multiple regression analysis was performed for each tooth region with the gingival thickness at the level of sulcus bottom, the width of keratinized gingiva, and gingival angle as the dependent variables. As the results, the NL group of the upper anterior teeth displayed, higher papilla height, and narrower keratinized gingiva, more acute gingival angle resulting in pronounced "scalloped" contour of the gingival margin, compared to the WS group. There was no significant difference between groups NL and WS with respect to probing depth and the gingival thickness. The regression analyses demonstrated that the gingival thickness in central incisors was significantly associated to the mesio-distal width and bucco-lingual width of the crown, and labial probing depth. The width of keratinized gingiva was significantly associated with labial probing depth in central incisors and with proximal probing depth and gingival angle in lateral incisors, and with labial and proximal probing depth, and gingival angle in canines. The gingival angle was significantly associated with papillary height and CW/CL ratio and additionally with proximal probing depth in central incisors, with the width of keratinized gingiva in lateral incisors, and with labial probing depth and the width of keratinized gingiva in canines. These results indicate that the form of clinical crown in upper anterior region could influence the clinical feature of gingiva and the influencing factors might be different according to the tooth region.

• Journal of IKEEE
• /
• v.5 no.2 s.9
• /
• pp.190-200
• /
• 2001

This paper is addressed to welding defect feature vector selection and implementation method of welding defect classifier using fuzzy techniques. We compare IAV, zero-crossing number as time domain analysis, power spectrum coefficient as frequency domain, histogram as both domain for welding defect feature selection. We choose histogram as feature vector by graph analysis and find out that maximum frequent occurrence number and section of corresponding signal scale in relative histogram show obvious difference between normal welding and voiding with penetration depth defect. We implement a fuzzy welding defect classifier using these feature vector, test it to verify its effectiveness for 695 welding data frame which consist of 4000 sampled data. As result of test, correct classification rate is 92.96%. Lab experimental results show a effectiveness of fuzzy welding defect classifier using relative histogram for practical Laser welding monitoring system in industry.