• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.1
• /
• pp.127-132
• /
• 2011

In this paper, we propose a novel method of adult image classification by combining skin color regions and edges in an input image. The proposed method consists of four steps. In the first step, initial skin color regions are detected by logical AND operation of all skin color regions detected by the existing methods of skin color detection. In the second step, a skin color probability map is created by modeling the distribution of skin color in the initial regions. Then, a binary image is generated by using threshold value from the skin color probability map. In the third step, after using the binary image and edge information, we detect final skin color regions using a region growing method. In the final step, adult image classification is performed by support vector machine(SVM). To this end, a feature vector is extracted by combining the final skin color regions and neighboring edges of them. As experimental results, the proposed method improves performance of the adult image classification by 9.6%, compared to the existing method.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.12
• /
• pp.7302-7308
• /
• 2014

Detecting target objects robustly in natural environments is a difficult problem in the computer vision and image processing areas. This paper suggests a method of robustly detecting target objects in the environments where reflection exists. The suggested algorithm first captures scenes with a stereo camera and extracts the line and corner features representing the target objects. This method then eliminates the reflected features among the extracted ones using a homographic transform. Subsequently, the method robustly detects the target objects by clustering only real features. The experimental results showed that the suggested algorithm effectively detects the target objects in reflection environments rather than existing algorithms.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.1
• /
• pp.60-65
• /
• 2010

This paper describes the results of 2D mapping, feature detection and matching to create the surrounding environment in the mounted stereo camera on Mobile robot. Extract method of image's feature in real-time processing for quick operation uses the edge detection and Sum of Absolute Difference(SAD), stereo matching technique can be obtained through the correlation coefficient. To estimate the location of a mobile robot using ZigBee beacon and encoders mounted on the robot is estimated by Kalman filter. In addition, the merged gyro scope to measure compass is possible to generate map during mobile robot is moving. The Simultaneous Localization and Mapping (SLAM) of mobile robot technology with an intelligent robot can be applied efficiently in human life would be based.

• Korean Journal of Cognitive Science
• /
• v.8 no.4
• /
• pp.55-62
• /
• 1997

We proposed a method for locating of mouth using deformable templates, described by a parameterized template. An energy function is defined which links, edges, peaks, valleys in image intensity to corresponding properties of the template. The template deforms itself by altering its parameter values to minimize the energy function. The minimized energy function's parameter values can be used as descriptors for the feature. We propose a method for locating mouth fast, accurately by limiting a range of parameters' value and getting initial value of parameters' by preprocessing.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.4
• /
• pp.1-8
• /
• 2023

In this paper, we propose a black ice detection platform framework using Convolutional Neural Networks (CNNs). To overcome black ice problem, we introduce a real-time based early warning platform using CNN-based architecture, and furthermore, in order to enhance the accuracy of black ice detection, we apply a multi-scale dilation convolution feature fusion (MsDC-FF) technique. Then, we establish a specialized experimental platform by using a comprehensive dataset of thermal road black ice images for a training and evaluation purpose. Experimental results of a real-time black ice detection platform show the better performance of our proposed network model compared to conventional image segmentation models. Our proposed platform have achieved real-time segmentation of road black ice areas by deploying a road black ice area segmentation network on the edge device Jetson Nano devices. This approach in parallel using multi-scale dilated convolutions with different dilation rates had faster segmentation speeds due to its smaller model parameters. The proposed MsCD-FF Net(2) model had the fastest segmentation speed at 5.53 frame per second (FPS). Thereby encouraging safe driving for motorists and providing decision support for road surface management in the road traffic monitoring department.

• Journal of the Korean association of regional geographers
• /
• v.2 no.2
• /
• pp.183-196
• /
• 1996

The purpose of this paper is to explore the possibility of automatic extraction of line feature from Satellite image. The first part reviews the relationship between spatial filtering and cartographic interpretation. The second part describes the principal operations of high frequency filters and their properties, the third part presents the result of filtering application to the SPOT Panchromatic image of the Chinju city. Some experimental results are given here indicating the high feasibility of the filtering technique. The results of the paper is summarized as follows: Firstly the good all-purposes filter dose not exist. Certain laplacian filter and Frei-chen filter were very sensitive to the noise and could not detect line features in our case. Secondly, summary filters and some other filters do an excellent job of identifying edges around urban objects. With the filtered image added to the original image, the interpretation is more easy. Thirdly, Compass gradient masks may be used to perform two-dimensional, discrete differentiation directional edge enhancement, however, in our case, the line featuring was not satisfactory. In general, the wide masks detect the broad edges and narrow masks are used to detect the sharper discontinuities. But, in our case, the difference between the $3{\times}3$ and $7{\times}7$ kernel filters are not remarkable. It may be due to the good spatial resolution of Spot scene. The filtering effect depends on local circumstance. Band or kernel size selection must be also considered. For the skillful geographical interpretation, we need to take account the more subtle qualitative information.

• Transactions on Semiconductor Engineering
• /
• v.2 no.1
• /
• pp.17-20
• /
• 2024

This paper presents a novel lightweight object detection model tailored for low-powered edge devices, addressing the limitations of traditional resource-intensive computer vision models. Our proposed detector, inspired by the Single Shot Detector (SSD), employs a compact yet robust network design. Crucially, it integrates an 'enhancer block' that significantly boosts its efficiency in detecting smaller objects. The model comprises two primary components: the Light_Block for efficient feature extraction using Depth-wise and Pointwise Convolution layers, and the Enhancer_Block for enhanced detection of tiny objects. Trained from scratch on the Udacity Annotated Dataset with image dimensions of 300x480, our model eschews the need for pre-trained classification weights. Weighing only 5.5MB with approximately 0.43M parameters, our detector achieved a mean average precision (mAP) of 27.7% and processed at 140 FPS, outperforming conventional models in both precision and efficiency. This research underscores the potential of lightweight designs in advancing object detection for edge devices without compromising accuracy.

• The KIPS Transactions:PartB
• /
• v.9B no.3
• /
• pp.327-336
• /
• 2002

Texture analysis is an important technique in many image understanding areas, such as perception of surface, object, shape and depth. But the previous works are intend to the issue of only texture segment, that is not capable of acquiring recognition information. No unsupervised method is basased on the recognition of texture in image. we propose a novel approach for efficient texture image analysis that uses unsupervised learning schemes for the texture recognition. The self-organization neural network for multiple texture image identification is based on block-based clustering and merging. The texture features used are the angle and magnitude in orientation-field that might be different from the sample textures. In order to show the performance of the proposed system, After we have attempted to build a various texture images. The final segmentation is achieved by using efficient edge detection algorithm applying to block-based dilation. The experimental results show that the performance of the system Is very successful.

• Journal of Internet Computing and Services
• /
• v.6 no.2
• /
• pp.59-68
• /
• 2005

In this paper, for compression and transmission of 3D image, we propose an algorithm which executes 3D discrete cosine transform(DCT) for 3D images, hierarchically segments 3D blocks of an image in comparison with the original image and executes finite-state vector quantization(FSVQ) for each 3D block. Using 3D DCT coefficient feature, a 3D image is segmented hierarchically into large smooth blocks and small edge blocks, then the block hierarchy informations are transmitted. The codebooks are constructed for each hierarchical blocks respectively, the encoder transmits codeword index using FSVQ for reducing encoded bit with hierarchical segmentation information. The new algorithm suggested in this paper shows that the quality of Small Lobster and Head image increased by 1,91 dB and 1.47 dB respectively compared with those of HFSVQ.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.12
• /
• pp.2313-2327
• /
• 1994

As the communication environment evolves, there is an increasing need for multiresolution image coding. To meet this need, the entrophy constratined vector quantizer(ECVQ) for coding of image pyramids by spline wavelet transform is introduced in this paper. This paper proposes a new scheme for image compression taking into account psychovisual feature both in the space and frequency domains : this proposed method involves two steps. First we use spline wavelet transform in order to obtain a set of biorthogonal subclasses of images ; the original image is decomposed at different scale using a pyramidal algorithm architecture. The decomposition is along the vertical and horizontal directions and maintains constant the number of pixels required the image. Second, according to Shannon's rate distortion theory, the wavelet coefficients are vectored quantized using a multi-resolution ECVQ(entropy-constrained vector quantizer) codebook. The simulation results showed that the proposed method could achieve higher quality LENA image improved by about 2.0 dB than that of the ECVQ using other wavelet at 0.5 bpp and, by about 0.5 dB at 1.0 bpp, and reduce the block effect and the edge degradation.