• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.2
• /
• pp.31-38
• /
• 2023

This paper introduces a real-time face swapping system that enables video influencers to swap their faces with arbitrary generated face images of their choice. The system is implemented as a Django-based server that uses a REST request to communicate with the generative model,specifically the pretrained stable diffusion model. Once generated, the generated image is displayed on the front page so that the influencer can decide whether to use the generated face or not, by clicking on the accept button on the front page. If they choose to use it, both their face and the generated face are sent to the landmark extraction module to extract the landmarks, which are then used to swap the faces. To minimize the fluctuation of landmarks over time that can cause instability or jitter in the output, a temporal filtering step is added. Furthermore, to increase the processing speed the system works on a reduced set of the extracted landmarks.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.4
• /
• pp.11-19
• /
• 2021

An abnormal object refers to a person, an object, or a mechanical device that performs abnormal and unusual behavior and needs observation or supervision. In order to detect this through artificial intelligence algorithm without continuous human intervention, a method of observing the specificity of temporal features using optical flow technique is widely used. In this study, an abnormal situation is identified by learning an algorithm that translates an input image frame to an optical flow image using a Generative Adversarial Network (GAN). In particular, we propose a technique that improves the pre-processing process to exclude unnecessary outliers and the post-processing process to increase the accuracy of identification in the test dataset after learning to improve the performance of the model's abnormal behavior identification. UCSD Pedestrian and UMN Unusual Crowd Activity were used as training datasets to detect abnormal behavior. For the proposed method, the frame-level AUC 0.9450 and EER 0.1317 were shown in the UCSD Ped2 dataset, which shows performance improvement compared to the models in the previous studies.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.40 no.1
• /
• pp.1-9
• /
• 2003

The CNN-UM algorithm which performs the analog parallel subtraction of images has been developed and its application study to the moving target detection has been done. The CNN-UM is the state of the art computation architecture with high computational potential of analog parallel processing. It is one of the strong candidates for the next generation of computing system which fulfills requirement of the real-time image processing. One weakness of the CNN-UM is that its analog parallel processing function is not fully utilized for the inter frame processing. If two subsequent image frames are superimposed with opposite signs on identical capacitors for short time period, the analog subtraction between them is achieved. The Principle of such temporal inter-frame processing algorithm has been described and its mathematical analysis has been done. Practical usefulness of the proposed algorithm has also been verified through the application for moving target detection.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.6
• /
• pp.107-112
• /
• 2022

In an embedded system such as a drone, it is difficult to store, transfer and analyze the entire hyper-spectral image to a server in real time because it takes a lot of power and time. Therefore, the hyper-spectral image data is transmitted to the server through dimension reduction or compression pre-processing. Feature selection method are used to send only the bands for analysis purpose, and these algorithms usually take a lot of processing time depending on the size of the image, even though the efficiency is high. In this paper, by improving the temporal disadvantage of the band selection algorithm, the time taken 24 hours was reduced to around 60-180 seconds based on the 40000*682 image resolution of 8GB data, and the use of 7.6GB RAM was significantly reduced to 2.3GB using 45 out of 150 bands. However, in terms of pixel classification performance, more than 98% of analysis results were derived similarly to the previous one.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.27 no.2
• /
• pp.119-125
• /
• 2017

The mud flat in Korea is the geographical feature generated from the sediment of rivers of Korea and China and it is the important topography for pollution purification and fishing industry. The mud flat is difficult to access such that it requires the aerial survey for the high-resolution spatial information of the area. In this study we used drones instead of the conventional aerial and remote sensing approaches which have shortcomings of costs and revisit times. We carried out GPS-based control point survey, temporal image acquisition using drones, bundle adjustment, stereo image processing for DSM and ortho photo generation, followed by co-registration between the spatio-temporal information.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.6C
• /
• pp.842-847
• /
• 2004

We present a new video coding technique to tradeoff frame rate and picture quality for low bit rate video coding. We show a model equation for selecting the optimal frame rate from the motion content of the source video. We can determine DCT quantization parameter (QP) using the frame rate and bit rate. For objective video quality measurement we propose a simple and effective error measure for skipped frames. The proposed method enhances the video quality up to 2 ㏈ over the H.263 TMN5 encoder.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.12 no.5
• /
• pp.12-18
• /
• 2004

Transient natures of diesel sprays are often characterized with spray visualization, since it is a non-intrusive and straightforward technique to be applied. However, as injection pressure is increased higher than a thousand bar in a modern direct injection diesel engine, very fine temporal and spatial resolutions in the spray visualization are required while sprays become optically denser. Discussed in this paper are macroscopic and microscopic spray visualization techniques and an example of image processing process for efficient and consistent measurement of spray parameters. The injection rate measurement method based on hydraulic pulse principle was suggested as a way of estimating injection velocity for transient diesel sprays. The spray visualization and injection rate measurement techniques were applied to analyze transient diesel sprays from a common-rail injection system and found to be practically effective.

• Journal of Information Processing Systems
• /
• v.17 no.2
• /
• pp.337-351
• /
• 2021

Automatically recognizing facial expressions in video sequences is a challenging task because there is little direct correlation between facial features and subjective emotions in video. To overcome the problem, a video facial expression recognition method using spatiotemporal recurrent neural network and feature fusion is proposed. Firstly, the video is preprocessed. Then, the double-layer cascade structure is used to detect a face in a video image. In addition, two deep convolutional neural networks are used to extract the time-domain and airspace facial features in the video. The spatial convolutional neural network is used to extract the spatial information features from each frame of the static expression images in the video. The temporal convolutional neural network is used to extract the dynamic information features from the optical flow information from multiple frames of expression images in the video. A multiplication fusion is performed with the spatiotemporal features learned by the two deep convolutional neural networks. Finally, the fused features are input to the support vector machine to realize the facial expression classification task. The experimental results on cNTERFACE, RML, and AFEW6.0 datasets show that the recognition rates obtained by the proposed method are as high as 88.67%, 70.32%, and 63.84%, respectively. Comparative experiments show that the proposed method obtains higher recognition accuracy than other recently reported methods.

• Annual Conference of KIPS
• /
• 2009.04a
• /
• pp.35-38
• /
• 2009

The paper presents an effective method to detect fire in video surveillance and monitoring system. The main contribution of this work is that we successfully use the Hidden Markov Models in the process of detecting the fire with a few preprocessing steps. First, the moving pixels detected from image difference, the color values obtained from the fire flames, and their pixels clustering are applied to obtain the image regions labeled as fire candidates; secondly, utilizing massive training data, including fire videos and non-fire videos, creates the Hidden Markov Models of fire and non-fire, which are used to make the final decision that whether the frame of the real-time video has fire or not in both temporal and spatial analysis. Experimental results demonstrate that it is not only robust but also has a very low false alarm rate, furthermore, on the ground that the HMM training which takes up the most time of our whole procedure is off-line calculated, the real-time detection and alarm can be well implemented when compared with the other existing methods.

• Annual Conference of KIPS
• /
• 2016.10a
• /
• pp.685-686
• /
• 2016

엑스선 투시촬영장치와 같은 의료용 동영상의 잡음 제거에 있어서 시공간 잡음제거 알고리즘을 제안한다. 제안된 알고리즘에서는 무한 충격 응답 기반 시간 영역 잡음 제거 얄고리즘을 도입하여 움직임이 적은 영역에서는 자연스러운 잡음 제거가 가능하며 움직임이 많은 영역에서는 연산량 측면에서 효율성을 고려하여 지역적 평균 필터 기반 공간 잡음 제거 알고리즘을 적용하여 움직임에 의한 흐려짐 열화 현상을 최소화 하면서 잡음 제거를 수행하였다.