• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.556-570
• /
• 2021

The existing video expression recognition methods mainly focus on the spatial feature extraction of video expression images, but tend to ignore the dynamic features of video sequences. To solve this problem, a multi-mode convolution neural network method is proposed to effectively improve the performance of facial expression recognition in video. Firstly, OpenFace 2.0 is used to detect face images in video, and two deep convolution neural networks are used to extract spatiotemporal expression features. Furthermore, spatial convolution neural network is used to extract the spatial information features of each static expression image, and the dynamic information feature is extracted from the optical flow information of multiple expression images based on temporal convolution neural network. Then, the spatiotemporal features learned by the two deep convolution neural networks are fused by multiplication. Finally, the fused features are input into support vector machine to realize the facial expression classification. Experimental results show that the recognition accuracy of the proposed method can reach 64.57% and 60.89%, respectively on RML and Baum-ls datasets. It is better than that of other contrast methods.

• Annual Conference of KIPS
• /
• 2021.11a
• /
• pp.704-707
• /
• 2021

Pigs' aggressive behavior represents one of the common issues that occur inside pigpens and which harm pigs' health and welfare, resulting in a financial burden to farmers. Continuously monitoring several pigs for 24 hours to identify those behaviors manually is a very difficult task for pig caretakers. In this study, we propose a lightweight video-based approach for monitoring pigs' aggressive behavior that can be implemented even in small-scale farms. The proposed system receives sequences of frames extracted from an RGB video stream containing pigs and uses MnasNet with a DM value of 0.5 to extract image features from pigs' ROI identified by predefined annotations. These extracted features are then forwarded to a lightweight LSTM to learn temporal features and perform behavior recognition. The experimental results show that our proposed model achieved 0.92 in recall and F1-score with an execution time of 118.16 ms/sequence.

• Journal of KIISE:Computing Practices and Letters
• /
• v.10 no.4
• /
• pp.309-318
• /
• 2004

This paper proposes the H.264-based selective fine granular scalable (FGS) coding scheme that selectively uses the temporal prediction data in the enhancement layer. The base layer of the proposed scheme is basically coded by the H.264 (MPEG-4 Part 10 AVC) visual coding scheme that is the state-of-art in codig efficiency. The enhancement layer is basically coded by the same bitplane-based algorithm of the MPEG-4 (Part 2) fine granular scalable coding scheme. In this paper, we introduce a new algorithm that uses the temproal prediction mechanism inside the enhancement layer and the effective selection mechanism to decide whether the temporally-predicted data would be sent to the decoder or not. Whenever applying the temporal prediction inside the enhancement layer, the temporal redundancies may be effectively reduced, however the drift problem would be severly occurred especially at the low bitrate transmission, due to the mismatch bewteen the encoder's and decoder's reference frame images. Proposed algorithm selectively uses the temporal-prediction data inside the enhancement layer only in case those data could siginificantly reduce the temporal redundancies, to minimize the drift error and thus to improve the overall coding efficiency. Simulation results, based on several test image sequences, show that the proposed scheme has 1∼3 dB higher coding efficiency than the H.264-based FGS coding scheme, even 3∼5 dB higher coding efficiency than the MPEG-4 FGS international standard.

• Journal of Korea Game Society
• /
• v.9 no.4
• /
• pp.89-96
• /
• 2009

In general, the size of volume data is large since it has logical 3D structure so it takes long time to manipulate. Much work has been done to improve processing speed of volume data. In this paper, we propose a interlaced scanning volume rendering that reduce computation time by using temporal coherence with minimum loss of image quality. It renders a current frame by reusing information of previous frame. Conventional volume raycasting renders each frame by casting rays on every pixels. On the other hand, our methods divided an image into n-pixel blocks, then it casts a ray on a pixel of a block per each frames. Consequently, it generates an image by accumulating pixel values of previous n frames. The quality of rendered image of our method is better than that of simple screen space subsampling method since it uses afterimage effect of human cognitive system, and it is n-times faster that the previous one.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.6
• /
• pp.1197-1206
• /
• 2019

Suncheon Bay is a very important natural resource and various efforts have been made to protect it from the environmental pollution. Although the project to monitor the environmental changes in periodically by observing the creatures in tidal flats is processing, it is being done inefficiently by people directly observing it. In this paper, we propose an object segmentation method that can be applied to the method to automatically monitor the living creatures in the tidal flats. In the proposed method, a foreground map representing the location of objects is obtained by using a temporal difference method, and a superpixel method is applied to detect the detailed boundary of an image. Finally the region of crab is extracted by combining the foreground map and the superpixel information. Experimental results show that the proposed method separates crab regions from a tidal flat image easily and accurately.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.8 no.1
• /
• pp.13-22
• /
• 2005

This paper classifies the topography of the Saemangeum Tidal flats based on Landsat TM satellite images by unsupervised ISODATA method, and analysis of the spatiotemporal changes of the classified shapes. The sedimental topography represents various properties according to the Saemangeum Tidal Embankment progress. We well proceed this study of the sedimental changes and distributions. By specifying the topographic characteristics of inner sea areas respectively, the investigation on the case study area according to the changes of the tidal will be useful in the establishment of land reclamation plan and the land use of the reclaimed area. In addition, the estuary image can be divided into tidal flats and sea surfaces using the band 4, also the detailed topography using the band 5, respectively among Landsat TM 7 bands. This paper contributes to the efficient image processing of the spatiotemporal sedimental changes.

• Journal of Korean Society for Geospatial Information Science
• /
• v.20 no.3
• /
• pp.41-48
• /
• 2012

Using multi-sensor or multi-temporal high resolution satellite images together is essential for efficient applications in remote sensing area. The purpose of this paper is to estimate geometric difference of translations between high-resolution optical and SAR images automatically. The geometric and radiometric pre-processing steps were fulfilled to calculate the similarity between optical and SAR images by using Mutual Information method. The coarsest-level pyramid images of each sensor constructed by gaussian pyramid method were generated to estimate the initial translation difference of the x, y directions for calculation efficiency. The precise geometric difference of translations was able to be estimated by applying this method from coarsest-level pyramid image to original image in order. Yet even when considered only translation between optical and SAR images, the proposed method showed RMSE lower than 5m in all study sites.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.22 no.10
• /
• pp.1327-1333
• /
• 2018

Numerical analysis has the advantage that no actual flow pathways need to be formulated, making this technique especially useful for simulation analysis such as pathway design. However, it does require that the complete physical parameters of the fluid and the complete boundary conditions be known. If any of them are unknown, either the calculation will become impossible, or even if the calculation does converge, the reliability of the result will be low. Therefore, a means of more accurate acquisition of flow information is required. In this paper, we present techniques for estimating flow field from a constraint equation for image information and velocity field, based on the image intensity changes accompanying the motion of dye in waterway. In the equation, we entered a stabilizing term to suppress estimation error. We show the effectiveness of our method through experiments with generated and real images of a Karman vortex.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.1
• /
• pp.55-61
• /
• 2020

With the development of deep learning technology and advances in computing power, video-based research is now gaining more and more attention. Video data contains a large amount of temporal and spatial information, which is the biggest difference compared with image data. It has a larger amount of data. It has attracted intense attention in computer vision. Among them, motion recognition is one of the research focuses. However, the action recognition of human in the video is extremely complex and challenging subject. Based on many research in human beings, we have found that artificial intelligence-like attention mechanisms are an efficient model for cognition. This efficient model is ideal for processing image information and complex continuous video information. We introduce this attention mechanism into video action recognition, paying attention to human actions in video and effectively improving recognition efficiency. In this paper, we propose a new 3D residual attention network using convolutional neural network based on two attention models to identify human action behavior in the video. An evaluation result of our model showed up to 90.7% accuracy.

• Journal of the Korean Geographical Society
• /
• v.50 no.5
• /
• pp.473-483
• /
• 2015

In a coastal area, a plenty of research has adopted remotely sensed data. This is because longterm interaction between land and ocean makes continuous geographical changes in a broad extent and unaccessible areas. However, conventional remote sensing platforms such as satellite or airplane has several disadvantages including limited temporal resolution and high operational costs. Hence, this study uses a UAV system to detect a coastline and its movement. Result of coastline detection shows how the coastline moves in a day. Time-series coastlines were derived from UAV aerial images through digital image processing. There is a drawback in the stability of UAV compared to the conventional remote sensing platform, but the advantage appears on the economical efficiency. Since the latest studies shows an improvement of UAV for a variety of purposes in many fields, a UAV can also be utilized for regional study and spatial data acquisition platform. geography can also utilize a UAV as a spatial data acquisition platform for regional study.