• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.8C
• /
• pp.505-515
• /
• 2011

This paper describes a normalization method of Haar-like features used for object detection. Previous method which performs variance normalization on Haar-like features requires a lot of calculations, since it uses an additional integral image for calculating the standard deviation of intensities of pixels in a candidate window and increases possibility of false detection in the area where variance of brightness is small. The proposed normalization method can be performed much faster than the previous method by not using additional integral image and classifiers which are trained with the proposed normalization method show robust performance in various lighting conditions. Experimental result shows that the object detector which uses the proposed method is 26% faster than the one which uses the previous method. Detection rate is also improved by 5% without increasing false alarm rate and 45% for the samples whose brightness varies significantly.

• Journal of KIISE:Databases
• /
• v.36 no.6
• /
• pp.446-454
• /
• 2009

Among various region detector and shape feature extraction method, MSER(Maximally Stable Extremal Region) and SIFT and its variant methods are popularly used in computer vision application. However, since SIFT is sensitive to the illumination change and MSER is sensitive to the scale change, it is not easy to apply the image similarity retrieval. In this paper, we present a Scale and Rotation Invariant Region Feature(SRIRF) descriptor using scale pyramid, MSER and affine normalization. The proposed SRIRF method is robust to scale, rotation, illumination change of image since it uses the affine normalization and the scale pyramid. We have tested the SRIRF method on various images. Experimental results demonstrate that the retrieval performance of the SRIRF method is about 20%, 38%, 11%, 24% better than those of traditional SIFT, PCA-SIFT, CE-SIFT and SURF, respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.12
• /
• pp.4439-4455
• /
• 2021

Compared with two-stage object detection algorithms, one-stage algorithms provide a better trade-off between real-time performance and accuracy. However, these methods treat the intermediate features equally, which lacks the flexibility to emphasize meaningful information for classification and location. Besides, they ignore the interaction of contextual information from different scales, which is important for medium and small objects detection. To tackle these problems, we propose an image pyramid network based on dual attention mechanism (DAIPNet), which builds an image pyramid to enrich the spatial information while emphasizing multi-scale informative features based on dual attention mechanisms for one-stage object detection. Our framework utilizes a pre-trained backbone as standard detection network, where the designed image pyramid network (IPN) is used as auxiliary network to provide complementary information. Here, the dual attention mechanism is composed of the adaptive feature fusion module (AFFM) and the progressive attention fusion module (PAFM). AFFM is designed to automatically pay attention to the feature maps with different importance from the backbone and auxiliary network, while PAFM is utilized to adaptively learn the channel attentive information in the context transfer process. Furthermore, in the IPN, we build an image pyramid to extract scale-wise features from downsampled images of different scales, where the features are further fused at different states to enrich scale-wise information and learn more comprehensive feature representations. Experimental results are shown on MS COCO dataset. Our proposed detector with a 300 × 300 input achieves superior performance of 32.6% mAP on the MS COCO test-dev compared with state-of-the-art methods.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.05a
• /
• pp.155-160
• /
• 2003

This paper presents a multiple face detector based on a robust pupil detection technique. The pupil detector uses active illumination that exploits the retro-reflectivity property of eyes to facilitate detection. The detection range of this method is appropriate for interactive desktop and kiosk applications. Once the location of the pupil candidates are computed, the candidates are filtered and grouped into pairs that correspond to faces using heuristic rules. To demonstrate the robustness of the face detection technique, a dual mode face tracker was developed, which is initialized with the most salient detected face. Recursive estimators are used to guarantee the stability of the process and combine the measurements from the multi-face detector and a feature correlation tracker. The estimated position of the face is used to control a pan-tilt servo mechanism in real-time, that moves the camera to keep the tracked face always centered in the image.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.8
• /
• pp.1823-1830
• /
• 2003

This paper presents a multiple face detector based on a robust pupil detection technique. The pupil detector uses active illumination that exploits the retro­reflectivity property of eyes to facilitate detection. The detection range of this method is appropriate for interactive desktop and kiosk applications. Once the location of the pupil candidates are computed, the candidates are filtered and grouped into pairs that correspond to faces using heuristic rules. To demonstrate the robustness of the face detection technique, a dual mode face tracker was developed, which is initialized with the most salient detected face. Recursive estimators are used to guarantee the stability of the process and combine the measurements from the multi­face detector and a feature correlation tracker. The estimated position of the face is used to control a pan­tilt servo mechanism in real­time, that moves the camera to keep the tracked face always centered in the image.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.49-52
• /
• 2019

Braille guide block of sidewalk is an essential facility for independent walking of the blind. The blind walks while checking the braile guide blocks with white cane and sense of sole. When they leave the braile area, they face difficulties until they find the braile guide blocks again. In this paper, we propose an algorithm that guides the walking of the blind by determining whether they follows the braille guide blocks safely. For this purpose, the slope of the braille block is selected as a feature and a 3-line detector is introduced. Also the slopes are stabilized using spatial filtering to deal with breaks or junctions of the braille block during the progress and temporal filtering to cope with ego-motion of the blind. Through simulations using a dataset obtained from the real sidewalks and indoors, it can be shown that the proposed algorithm can successfully estimate the walking direction and determine whether the blind is out of the braille guide block area.

• Proceedings of the Korea Crystallographic Association Conference
• /
• 2002.11a
• /
• pp.5-6
• /
• 2002

The new spectrometer for X-ray Induced Electron Emission Spectroscopy (XIEES) .has been recently developed in KRISS in collaboration with PTI (Russia). The spectrometer allows to perform research using the XAFS, SXAFS, XANES techniques (D.C.Koningsberger and R.Prins, 1988) as well as the number of techniques from XIEES field(L.A.Bakaleinikov et all, 1992). The experiments may be carried out with registration of transmitted through the sample x-rays (to investigate bulk samples) or/and total electron yield (TEY) from the sample surface that gives the high (down to several atomic mono-layers in soft x-ray region) near surface sensitivity. The combination of these methods together give the possibility to obtain a quantitative information on elemental composition, chemical state, atomic structure for powder samples and solids, including non-crystalline materials (the long range order is not required). The optical design of spectrometer is made according to Johannesson true focusing schematics and presented on the Fig.1. Five stepping motors are used to maintain the focusing condition during the photon energy scan (crystal angle, crystal position along rail, sample goniometer rail angle, sample goniometer position along rail and sample goniometer angle relatively of rail). All movements can be done independently and simultaneously that speeds up the setting of photon energy and allows the using of crystals with different Rowland radil. At present six curved crystals with different d-values and one flat synthetic multilayer are installed on revolver-type monochromator. This arrangement allows the wide range of x-rays from 100 eV up to 25 keV to be obtained. Another 4 stepping motors set exit slit width, sample angle, channeltron position and x-ray detector position. The differential pumping allows to unite vacuum chambers of spectrometer and x-ray generator avoiding the absorption of soft x-rays on Be foil of a window and in atmosphere. Another feature of vacuum system is separation of walls of vacuum chamber (which are deformed by the atmospheric pressure) from optical elements of spectrometer. This warrantees that the optical elements are precisely positioned. The detecting system of the spectrometer consists of two proportional counters, one scintillating detector and one channeltron detector. First proportional counter can be used as I/sub 0/-detector in transmission mode or by measuring the fluorescence from exit slit edge. The last installation can be used to measure the reference data (that is necessary in XANES measurements), in this case the reference sample is installed on slit knife edge. The second proportional counter measures the intensity of x-rays transmitted through the sample. The scintillating detector is used in the same way but on the air for the hard x-rays and for alignment purposes. Total electron yield from the sample is measured by channeltron. The spectrometer is fully controlled by special software that gives the high flexibility and reliability in carrying out of the experiments. Fig.2 and fig.3 present the typical XAFS spectra measured with spectrometer.

• Korean Journal of Remote Sensing
• /
• v.28 no.4
• /
• pp.357-367
• /
• 2012

This paper presents a two-stage methodology for anomaly detection from hyperspectral imagery that consists of transform-based feature extraction and selection, and computation of a local spatial auto-correlation statistic. First, principal component transform and 3D wavelet transform are applied to reduce redundant spectral information from hyperspectral imagery. Then feature selection based on global skewness and the portion of highly skewed sub-areas is followed to find optimal features for anomaly detection. Finally, a local indicator of spatial association (LISA) statistic is computed to account for both spectral and spatial information unlike traditional anomaly detection methodology based only on spectral information. An experiment using airborne CASI imagery is carried out to illustrate the applicability of the proposed anomaly detection methodology. From the experiments, anomaly detection based on the LISA statistic linked with the selection of optimal features outperformed both the traditional RX detector which uses only spectral information, and the case using major principal components with large eigen-values. The combination of low- and high-frequency components by 3D wavelet transform showed the best detection capability, compared with the case using optimal features selected from principal components.

• Asian-Australasian Journal of Animal Sciences
• /
• v.26 no.7
• /
• pp.1030-1037
• /
• 2013

Early detection of anomalies is an important issue in the management of group-housed livestock. In particular, failure to detect oestrus in a timely and accurate way can become a limiting factor in achieving efficient reproductive performance. Although a rich variety of methods has been introduced for the detection of oestrus, a more accurate and practical method is still required. In this paper, we propose an efficient data mining solution for the detection of oestrus, using the sound data of Korean native cows (Bos taurus coreanea). In this method, we extracted the mel frequency cepstrum coefficients from sound data with a feature dimension reduction, and use the support vector data description as an early anomaly detector. Our experimental results show that this method can be used to detect oestrus both economically (even a cheap microphone) and accurately (over 94% accuracy), either as a standalone solution or to complement known methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.11
• /
• pp.2720-2736
• /
• 2013

Multi-view face detection has become an active area for research in the last few years. In this paper, a novel multi-view human face detection algorithm based on improved real Adaboost is presented. Real Adaboost algorithm is improved by weighted combination of weak classifiers and the approximately best combination coefficients are obtained. After that, we proved that the function of sample weight adjusting method and weak classifier training method is to guarantee the independence of weak classifiers. A coarse-to-fine hierarchical face detector combining the high efficiency of Haar feature with pose estimation phase based on our real Adaboost algorithm is proposed. This algorithm reduces training time cost greatly compared with classical real Adaboost algorithm. In addition, it speeds up strong classifier converging and reduces the number of weak classifiers. For frontal face detection, the experiments on MIT+CMU frontal face test set result a 96.4% correct rate with 528 false alarms; for multi-view face in real time test set result a 94.7 % correct rate. The experimental results verified the effectiveness of the proposed approach.