• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2009.10a
• /
• pp.574-578
• /
• 2009

Real-time object detection for distinguishing a moving object of interests from the background image in still image or video image sequence is an essential step to a correct object tracking and recognition. Moving cast shadow can be misclassified as part of objects or moving objects because the shadow region is included in the moving object region after object segmentation. For this reason, an algorithm for shadow removal plays an important role in the results of accurate moving object detection and tracking systems. To handle with the problems, an accurate algorithm based on the features of moving object and shadow in color space is presented in this paper. Experimental results show that the proposed algorithm is effective to detect a moving object and to remove shadow in test video sequences.

• 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.

• Tunnel and Underground Space
• /
• v.33 no.4
• /
• pp.281-298
• /
• 2023

Continuous research efforts are being devoted to unmanned mobile platforms for lunar exploration. There is an ongoing demand for real-time information processing to accurately determine the positioning and mapping of areas of interest on the lunar surface. To apply deep learning processing and analysis techniques to practical rovers, research on software integration and optimization is imperative. In this study, a foundational investigation has been conducted on real-time analysis of virtual lunar base construction site images, aimed at automatically quantifying spatial information of key objects. This study involved transitioning from an existing region-based object recognition algorithm to a boundary box-based algorithm, thus enhancing object recognition accuracy and inference speed. To facilitate extensive data-based object matching training, the Batch Hard Triplet Mining technique was introduced, and research was conducted to optimize both training and inference processes. Furthermore, an improved software system for object recognition and identical object matching was integrated, accompanied by the development of visualization software for the automatic matching of identical objects within input images. Leveraging satellite simulative captured video data for training objects and moving object-captured video data for inference, training and inference for identical object matching were successfully executed. The outcomes of this research suggest the feasibility of implementing 3D spatial information based on continuous-capture video data of mobile platforms and utilizing it for positioning objects within regions of interest. As a result, these findings are expected to contribute to the integration of an automated on-site system for video-based construction monitoring and control of significant target objects within future lunar base construction sites.

• Smart Media Journal
• /
• v.13 no.2
• /
• pp.68-84
• /
• 2024

This paper describes the development of a system and algorithms for high-quality welfare services by recognizing behavior development indicators (activity, sociability, danger) in children aged 0 to 2 years old using action recognition technology. Action recognition targeted 11 behaviors from lying down in 0-year-olds to jumping in 2-year-olds, using data directly obtained from actual videos provided for research purposes by three nurseries in the Gwangju and Jeonnam regions. A dataset of 1,867 actions from 425 clip videos was built for these 11 behaviors, achieving an average recognition accuracy of 97.4%. Additionally, for real-world application, the Edge Video Analyzer (EVA), a behavior analysis device, was developed and implemented with a region-specific random frame selection-based PoseC3D algorithm, capable of recognizing actions in real-time for up to 30 people in four-channel videos. The developed system was installed in three nurseries, tested by ten childcare teachers over a month, and evaluated through surveys, resulting in a perceived accuracy of 91 points and a service satisfaction score of 94 points.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.4
• /
• pp.189-198
• /
• 2023

As crimes frequently occur on the street, the spread of CCTV is increasing. However, due to the shortcomings of passively operated CCTV, the need for intelligent CCTV is attracting attention. Due to the heavy system of such intelligent CCTV, high-performance devices are required, which has a problem in that it is expensive to replace the general CCTV. To solve this problem, an intelligent CCTV system that recognizes low-quality images and operates even on devices with low performance is required. Therefore, this paper proposes a Saying CCTV system that can detect threats in real time by using the AWS cloud platform to lighten the system and convert images into text. Based on the data extracted using YOLO v4 and OpenPose, it is implemented to determine the risk object, threat behavior, and threat situation, and calculate the risk using machine learning. Through this, the system can be operated anytime and anywhere as long as the network is connected, and the system can be used even with devices with minimal performance for video shooting and image upload. Furthermore, it is possible to quickly prevent crime by automating meaningful statistics on crime by analyzing the video and using the data stored as text.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.5
• /
• pp.1103-1109
• /
• 2014

The motion recognition system has been broadly studied in digital image and video processing fields. Recently, method using th depth image is used very useful. However, recognition accuracy of depth image based method will be loss caused by size and shape of object distorted for angle of the depth sensor. Therefore, distortion correction of depth sensor is positively necessary for distinguished performance of the recognition system. In this paper, we propose a pre-processing algorithm to improve the motion recognition system. Depth data from depth sensor converted to real world, performed the corrected angle, and then inverse converted to projective world. The proposed system make progress using the OpenCV and the window program, and we test a system using the Kinect in real time. In addition, designed using Verilog-HDL and verified through the Zynq-7000 FPGA Board of Xilinx.

• Journal of Internet Computing and Services
• /
• v.8 no.3
• /
• pp.119-129
• /
• 2007

To recognition a human motion, in this paper, we propose a neural approach using silhouettes in video frames captured by two cameras placed at the front and side of the human body. To extract features of the silhouettes for motion estimation, the proposed system computes both global and local features and then groups these features into static and dynamic features depending on whether features are in a static frame. Extracted features are in a static frame. Extracted features are used to train a RBF network. The neural system uses static features as the input of the neural network and dynamic features as additional features for recognition. In this paper, the proposed method was applied to movement education for young children. The basic movements for such education consist of locomotor movements, such as walking, jumping, and hopping, and non-locomotor movements, including bending, stretching, balancing and turning. The system demonstrated the effectiveness of motion recognition for movement education generated by the proposed neural network. The proposed system dan be extended to the system for movement education which develops the spatial sense of young children.

• Journal of Advanced Navigation Technology
• /
• v.24 no.2
• /
• pp.155-162
• /
• 2020

Recently, with the development of computer vision and deep learning technology, research on human action recognition has been actively conducted for video analysis, video surveillance, interactive multimedia, and human machine interaction applications. Diverse techniques have been introduced for human action understanding and classification by many researchers using RGB image, depth image, skeleton and inertial data. However, skeleton-based action discrimination is still a challenging research topic for human machine-interaction. In this paper, we propose an end-to-end skeleton joints mapping of action for generating spatio-temporal image so-called dynamic image. Then, an efficient deep convolution neural network is devised to perform the classification among the action classes. We use publicly accessible UTD-MHAD skeleton dataset for evaluating the performance of the proposed method. As a result of the experiment, the proposed system shows better performance than the existing methods with high accuracy of 97.45%.

• Annual Conference of KIPS
• /
• 2021.05a
• /
• pp.354-357
• /
• 2021

Infants communicate their feelings and needs to the outside world through non-verbal methods such as crying and displaying diverse facial expressions. However, inexperienced parents tend to decode these non-verbal messages incorrectly and take inappropriate actions, which might affect the bonding they build with their babies and the cognitive development of the newborns. In this paper, we propose an aural-visual two-stream based infant cry recognition system to help parents comprehend the feelings and needs of crying babies. The proposed system first extracts the features from the pre-processed audio and video data by using the VGGish model and 3D-CNN model respectively, fuses the extracted features using a fully connected layer, and finally applies a SoftMax function to classify the fused features and recognize the corresponding type of cry. The experimental results show that the proposed system classification exceeds 0.92 in F1-score, which is 0.08 and 0.10 higher than the single-stream aural model and single-stream visual model.

• Journal of Korea Multimedia Society
• /
• v.19 no.2
• /
• pp.252-263
• /
• 2016

Road sign is an important street furniture that gives some information such as road conditions, driving direction and condition for a driver. Thus, road sign is a major target of image recognition for self-driving car, ADAS(autonomous vehicle and intelligent driver assistance systems), and ITS(intelligent transport systems). In this paper, an enhanced road sign recognition system is proposed for MMS(Mobile Mapping System) using the single camera and GPS. For the proposed system, first, a road sign recognition scheme is proposed. this scheme is composed of detection and classification step. In the detection step, object candidate regions are extracted in image frames using hybrid road sign detection scheme that is based on color and shape features of road signs. And, in the classification step, the area of candidate regions and road sign template are compared. Second, a Geo-marking scheme for geo-content that is consist of road sign image and coordinate value is proposed. If the serious situation such as car accident is happened, this scheme can protect geographical information of road sign against illegal users. By experiments with test video set, in the three parts that are road sign recognition, coordinate value estimation and geo-marking, it is confirmed that proposed schemes can be used for MMS in commercial area.