• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.9
• /
• pp.1649-1656
• /
• 2016

Performance of the existing sensor-based fire detection system is limited according to factors in the environment surrounding the sensor. A number of image-based fire detection systems were introduced in order to solve these problem. But such a system can generate a false alarm for objects similar in appearance to fire due to algorithm that directly defines the characteristics of a flame. Also fir detection systems using movement between video flames cannot operate correctly as intended in an environment in which the network is unstable. In this paper, we propose an image-based fire detection method using CNN (Convolutional Neural Network). In this method, firstly we extract fire candidate region using color information from video frame input and then detect fire using trained CNN. Also, we show that the performance is significantly improved compared to the detection rate and missing rate found in previous studies.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.14 no.5
• /
• pp.2430-2436
• /
• 2013

This study suggests a method of detecting the still object, which becomes a cause of danger within the crossroad. The Inverse Perspective Transform was performed in order to make the object size consistent by being inputted the real-time image from CCTV that is installed within the crossroad. It established the detection area in the image with the perspective transform and generated the adaptative background image with the use of the moving information on object. The detection of the stop object was detected the candidate region of the stop object by using the background-image differential method. To grasp the appearance of truth on the detected candidate region, a method is proposed that uses the gradient information on image and EHD(Edge Histogram Descriptor). To examine performance of the suggested algorithm, it experimented by storing the images in the commuting time and the daytime through DVR, which is installed on the cross street. As a result of experiment, it could efficiently detect the stop vehicle within the detection region inside the crossroad. The processing speed is shown in 13~18 frame per second according to the area of the detection region, thereby being judged to likely have no problem about the real-time processing.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.7
• /
• pp.57-64
• /
• 2020

In this paper, we propose an scalable re-detection for correlation filter in visual tracking. In real world, there are lots of target disappearances and reappearances during tracking, thus failure detection and re-detection methods are needed. One of the important point for re-detection is that a searching area must be large enough to find the missing target. For robust visual tracking, we adopt kernelized correlation filter as a baseline. Correlation filters have been extensively studied for visual object tracking in recent years. However conventional correlation filters detect the target in the same size area with the trained filter which is only 2 to 3 times larger than the target. When the target is disappeared for a long time, we need to search a wide area to re-detect the target. Proposed algorithm can search the target in a scalable area, hence the searching area is expanded by 2% in every frame from the target loss. Four datasets are used for experiments and both qualitative and quantitative results are shown in this paper. Our algorithm succeed the target re-detection in challenging datasets while conventional correlation filter fails.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.11
• /
• pp.5522-5536
• /
• 2017

In this paper, a video based smoke detection method using dynamic texture feature extraction with volume local binary patterns is studied. Block based method was used to distinguish smoke frames in high definition videos obtained by experiments firstly. Then we propose a method that directly extracts dynamic texture features based on irregular motion regions to reduce adverse impacts of block size and motion area ratio threshold. Several general volume local binary patterns were used to extract dynamic texture, including LBPTOP, VLBP, CLBPTOP and CVLBP, to study the effect of the number of sample points, frame interval and modes of the operator on smoke detection. Support vector machine was used as the classifier for dynamic texture features. The results show that dynamic texture is a reliable clue for video based smoke detection. It is generally conducive to reducing the false alarm rate by increasing the dimension of the feature vector. However, it does not always contribute to the improvement of the detection rate. Additionally, it is found that the feature computing time is not directly related to the vector dimension in our experiments, which is important for the realization of real-time detection.

• Journal of the Korea Society of Computer and Information
• /
• v.15 no.5
• /
• pp.57-65
• /
• 2010

In this paper, we propose a new real-time algorithm detecting the flame and smoke in digital CCTV images. Because the forest fire causes the enormous human life and damage of property, the early management according to the early sensing is very important. The proposed algorithm for monitoring forest fire is classified into the flame sensing and detection of smoke. The flame sensing algorithm detects a flame through the conditional test at YCbCr color model from the single frame. For the detection of smoke, firstly the background range is set by using differences between current picture and the average picture among the adjacent frames in the weighted value, and the pixels which get out of this range and have a gray-scale are detected in the smoke area. Because the proposed flame sensing algorithm is stronger than the existing algorithms in the change of the illuminance according to the quantity of sunshine, and the smoke detection algorithm senses the pixel of a gray-scale with the smoke considering the amount of change for unit time, the effective early forest fire detection is possible. The experimental results indicate that the proposed algorithm provides better performance than existing algorithms.

• Smart Structures and Systems
• /
• v.30 no.6
• /
• pp.613-626
• /
• 2022

Guaranteeing the quality and integrity of structural health monitoring (SHM) data is very important for an effective assessment of structural condition. However, sensory system may malfunction due to sensor fault or harsh operational environment, resulting in multiple types of data anomaly existing in the measured data. Efficiently and automatically identifying anomalies from the vast amounts of measured data is significant for assessing the structural conditions and early warning for structural failure in SHM. The major challenges of current automated data anomaly detection methods are the imbalance of dataset categories. In terms of the feature of actual anomalous data, this paper proposes a data anomaly detection method based on data-level and deep learning technique for SHM of civil engineering structures. The proposed method consists of a data balancing phase to prepare a comprehensive training dataset based on data-level technique, and an anomaly detection phase based on a sophisticatedly designed network. The advanced densely connected convolutional network (DenseNet) and Transformer encoder are embedded in the specific network to facilitate extraction of both detail and global features of response data, and to establish the mapping between the highest level of abstractive features and data anomaly class. Numerical studies on a steel frame model are conducted to evaluate the performance and noise immunity of using the proposed network for data anomaly detection. The applicability of the proposed method for data anomaly classification is validated with the measured data of a practical supertall structure. The proposed method presents a remarkable performance on data anomaly detection, which reaches a 95.7% overall accuracy with practical engineering structural monitoring data, which demonstrates the effectiveness of data balancing and the robust classification capability of the proposed network.

• Science of Emotion and Sensibility
• /
• v.20 no.2
• /
• pp.137-148
• /
• 2017

The present study examined the effect of stimulus contrast and color on detection of a target embedded in streams of letters. In Experiment 1, each trial displayed four rapid serial visual presentation (RSVP) streams of letters (i.e., multi-RSVP), and each stream occupied one of four different locations. Each frame in the RSVP stream had four white distractors at the locations except one frame where a dim grey target was displayed at a location with three white distractors at the remaining locations. In the low-visibility target condition, the target's grey color was slightly darker than the background grey whereas much dimmer in the high-visibility condition. Participants were asked to report presence of a predesignated target as quickly and accurately as possible upon its detection in each trial, and their target detection turned out more accurate and quicker in the high-visibility than the low-visibility condition. In Experiment 2, the same RSVP displays and task as Experiment were used, but the grey target letters in the high-visibility condition were replaced with those of distinct chromatic colors. Participants detected target presence more accurately in the high-visibility condition, but the reaction time did not differ between the visibility conditions. The results indicate that higher stimulus contrast as well as distinct color can improve perception of a target stimulus displayed among visually-demanding background, but also suggest that stimulus contrast may play a more substantial role for such perceptual improvement.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.7
• /
• pp.1399-1404
• /
• 2004

The development of computer technology and the advancement of the technology of information and communications spread the technology of multimedia and increased the use of multimedia data with large capacity. Users can grasp the overall video data and they are able to play wanted video back. To grasp the overall video data it is necessary to offer the list of summarized video data information. In order to search video efficiently an index process of video data is essential and it is also indispensable skill. Therefore, this thesis suggested the effective method about the cut detection of frames which will become a basis of an index based on contents of video image data. This suggested method was detected as the unchanging pixel rotor intelligence value, classified into diagonal direction. Pixel value of color detected in each frame of video data is stored as A(i, i) matrix - i is the number of frames, i is an image height of frame. By using the stored pixel value as the method of UE(Mean Square Error) I calculated a specified value difference between frames and detected cut quickly and exactly in case it is bigger than threshold value set in advance. To carry out an experiment on the cut detection of lames comprehensively, 1 experimented on many kinds of video, analyzing and comparing efficiency of the cut detection system.

• Journal of Intelligence and Information Systems
• /
• v.28 no.1
• /
• pp.89-106
• /
• 2022

Recently, the amount of video data collected from smartphones, CCTVs, black boxes, and high-definition cameras has increased rapidly. According to the increasing video data, the requirements for analysis and utilization are increasing. Due to the lack of skilled manpower to analyze videos in many industries, machine learning and artificial intelligence are actively used to assist manpower. In this situation, the demand for various computer vision technologies such as object detection and tracking, action detection, emotion detection, and Re-ID also increased rapidly. However, the object detection and tracking technology has many difficulties that degrade performance, such as re-appearance after the object's departure from the video recording location, and occlusion. Accordingly, action and emotion detection models based on object detection and tracking models also have difficulties in extracting data for each object. In addition, deep learning architectures consist of various models suffer from performance degradation due to bottlenects and lack of optimization. In this study, we propose an video analysis system consists of YOLOv5 based DeepSORT object tracking model, SlowFast based action recognition model, Torchreid based Re-ID model, and AWS Rekognition which is emotion recognition service. Proposed model uses single-linkage hierarchical clustering based Re-ID and some processing method which maximize hardware throughput. It has higher accuracy than the performance of the re-identification model using simple metrics, near real-time processing performance, and prevents tracking failure due to object departure and re-emergence, occlusion, etc. By continuously linking the action and facial emotion detection results of each object to the same object, it is possible to efficiently analyze videos. The re-identification model extracts a feature vector from the bounding box of object image detected by the object tracking model for each frame, and applies the single-linkage hierarchical clustering from the past frame using the extracted feature vectors to identify the same object that failed to track. Through the above process, it is possible to re-track the same object that has failed to tracking in the case of re-appearance or occlusion after leaving the video location. As a result, action and facial emotion detection results of the newly recognized object due to the tracking fails can be linked to those of the object that appeared in the past. On the other hand, as a way to improve processing performance, we introduce Bounding Box Queue by Object and Feature Queue method that can reduce RAM memory requirements while maximizing GPU memory throughput. Also we introduce the IoF(Intersection over Face) algorithm that allows facial emotion recognized through AWS Rekognition to be linked with object tracking information. The academic significance of this study is that the two-stage re-identification model can have real-time performance even in a high-cost environment that performs action and facial emotion detection according to processing techniques without reducing the accuracy by using simple metrics to achieve real-time performance. The practical implication of this study is that in various industrial fields that require action and facial emotion detection but have many difficulties due to the fails in object tracking can analyze videos effectively through proposed model. Proposed model which has high accuracy of retrace and processing performance can be used in various fields such as intelligent monitoring, observation services and behavioral or psychological analysis services where the integration of tracking information and extracted metadata creates greate industrial and business value. In the future, in order to measure the object tracking performance more precisely, there is a need to conduct an experiment using the MOT Challenge dataset, which is data used by many international conferences. We will investigate the problem that the IoF algorithm cannot solve to develop an additional complementary algorithm. In addition, we plan to conduct additional research to apply this model to various fields' dataset related to intelligent video analysis.

• Journal of Intelligence and Information Systems
• /
• v.28 no.4
• /
• pp.329-346
• /
• 2022

Recently, as the market for short form videos (Instagram, TikTok, YouTube) on social media has gradually increased, research using them is actively being conducted in the artificial intelligence field. A representative research field is Video to Shop, which detects fashion products in videos and searches for product images. In such a video-based artificial intelligence model, product features are extracted using convolution operations. However, due to the limitation of computational resources, extracting features using all the frames in the video is practically impossible. For this reason, existing studies have improved the model's performance by sampling only a part of the entire frame or developing a sampling method using the subject's characteristics. In the existing Video to Shop study, when sampling frames, some frames are randomly sampled or sampled at even intervals. However, this sampling method degrades the performance of the fashion product search model while sampling noise frames where the product does not exist. Therefore, this paper proposes a sampling method MF (Missing Fashion items on frame) sampler that removes noise frames and improves the performance of the search model. MF sampler has improved the problem of resource limitations by developing a keyframe mechanism. In addition, the performance of the search model is improved through noise frame removal using the noise detection model. As a result of the experiment, it was confirmed that the proposed method improves the model's performance and helps the model training to be effective.