• ETRI Journal
• /
• v.39 no.4
• /
• pp.502-513
• /
• 2017

For video understanding, namely analyzing who did what in a video, actions along with objects are primary elements. Most studies on actions have handled recognition problems for a well-trimmed video and focused on enhancing their classification performance. However, action detection, including localization as well as recognition, is required because, in general, actions intersect in time and space. In addition, most studies have not considered extensibility for a newly added action that has been previously trained. Therefore, proposed in this paper is an extensible hierarchical method for detecting generic actions, which combine object movements and spatial relations between two objects, and inherited actions, which are determined by the related objects through an ontology and rule based methodology. The hierarchical design of the method enables it to detect any interactive actions based on the spatial relations between two objects. The method using object information achieves an F-measure of 90.27%. Moreover, this paper describes the extensibility of the method for a new action contained in a video from a video domain that is different from the dataset used.

• Journal of Korea Multimedia Society
• /
• v.22 no.8
• /
• pp.823-831
• /
• 2019

In recent years, there has been developed systems such as a surveillance system and access control using a face recognition function instead of a password or an RFID chip, thereby reducing the risk of falsification. Moreover, deep learning technology has been applied to real-time face recognition technology in video, so it makes possible the development of access control system that improves the accuracy of recognition and efficiency of management. In this paper, we propose a real-time access management system based on face recognition using ResNet. The system is based on web server, which make it possible to manage the access by recognizing the person of the image through the camera and access information stored in the database. It can be accessed by a user application to receive various information. The implemented system identifies a person in real time and allows access control by accurately distinguishing whether they are members or not, and the test results can recognize in 0.2 seconds. The accuracy of recognition rate is up to about 97% depending on the experiment environment. With this system, access can be managed quickly and effectively, even many people rush to it.

• Journal of Multimedia Information System
• /
• v.9 no.2
• /
• pp.103-112
• /
• 2022

Background: A presentation attack places the printed image or displayed video at the front of the sensor to deceive the biometric recognition system. Usually, presentation attackers steal a genuine user's biometric image and use it for presentation attack. In recent years, reconstruction attack and adversarial attack can generate high-quality fake images, and have high attack success rates. However, their attack rates degrade remarkably after image shooting. Methods: In order to comprehensively analyze the threat of presentation attack to palmprint recognition system, this paper makes six palmprint presentation attack datasets. The datasets were tested on texture coding-based recognition methods and deep learning-based recognition methods. Results and conclusion: The experimental results show that the presentation attack caused by the leakage of the original image has a high success rate and a great threat; while the success rates of reconstruction attack and adversarial attack decrease significantly.

• Smart Media Journal
• /
• v.9 no.1
• /
• pp.23-29
• /
• 2020

This paper presents an approach for dynamic hand gesture recognition by using algorithm based on 3D Convolutional Neural Network (3D_CNN), which is later extended to 3D Residual Networks (3D_ResNet), and the neural network based key frame selection. Typically, 3D deep neural network is used to classify gestures from the input of image frames, randomly sampled from a video data. In this work, to improve the classification performance, we employ key frames which represent the overall video, as the input of the classification network. The key frames are extracted by SegNet instead of conventional clustering algorithms for video summarization (VSUMM) which require heavy computation. By using a deep neural network, key frame selection can be performed in a real-time system. Experiments are conducted using 3D convolutional kernels such as 3D_CNN, Inflated 3D_CNN (I3D) and 3D_ResNet for gesture classification. Our algorithm achieved up to 97.8% of classification accuracy on the Cambridge gesture dataset. The experimental results show that the proposed approach is efficient and outperforms existing methods.

• The Journal of Korea Robotics Society
• /
• v.1 no.2
• /
• pp.172-179
• /
• 2006

In this paper, we present a practical place and object recognition method for guiding visitors in building environments. Recognizing places or objects in real world can be a difficult problem due to motion blur and camera noise. In this work, we present a modeling method based on the bidirectional interaction between places and objects for simultaneous reinforcement for the robust recognition. The unification of visual context including scene context, object context, and temporal context is also. The proposed system has been tested to guide visitors in a large scale building environment (10 topological places, 80 3D objects).

• International journal of advanced smart convergence
• /
• v.9 no.1
• /
• pp.113-120
• /
• 2020

Recently, the license plate OCR system has been commercialized in a variety of fields and preferred utilizing low-cost embedded systems using only cameras. This system has a high recognition rate of about 98% or more for the environments such as parking lots where non-vehicle is restricted; however, the environments where non-vehicle objects are not restricted, the recognition rate is about 50% to 70%. This low performance is due to the changes in the environment by non-vehicle objects in real-time situations that occur anomaly data which is similar to the license plates. In this paper, we implement the appropriate anomaly detection based on semi-supervised learning for the license plate OCR system in the real-time environment where the appearance of non-vehicle objects is not restricted. In the experiment, we compare systems which anomaly detection is not implemented in the preceding research with the proposed system in this paper. As a result, the systems which anomaly detection is not implemented had a recognition rate of 77%; however, the systems with the semi-supervised learning based on anomaly detection had 88% of recognition rate. Using the techniques of anomaly detection based on the semi-supervised learning was effective in detecting anomaly data and it was helpful to improve the recognition rate of real-time situations.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.2
• /
• pp.181-190
• /
• 2009

This paper presents a multimodal emergency recognition system based on visual information, audio information and gravity sensor information. It consists of video processing module, audio processing module, gravity sensor processing module and multimodal integration module. The video processing module and gravity sensor processing module respectively detects actions such as moving, stopping and fainting and transfer them to the multimodal integration module. The multimodal integration module detects emergency by fusing the transferred information and verifies it by asking a question and recognizing the answer via audio channel. The experiment results show that the recognition rate of video processing module only is 91.5% and that of gravity sensor processing module only is 94%, but when both information are combined the recognition result becomes 100%.

• Journal of the Korea Computer Industry Society
• /
• v.3 no.7
• /
• pp.901-914
• /
• 2002

In this paper, We propose an efficient automatic caption detection and location method, caption recognition using FE-MCBP(Feature Extraction based Multichained BackPropagation) neural network for content based retrieval of video. Frames are selected at fixed time interval from video and key frames are selected by gray scale histogram method. for each key frames, segmentation is performed and caption lines are detected using line scan method. lastly each characters are separated. This research improves speed and efficiency by color segmentation using local maximum analysis method before line scanning. Caption detection is a first stage of multimedia database organization and detected captions are used as input of text recognition system. Recognized captions can be searched by content based retrieval method.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.3 no.2
• /
• pp.178-186
• /
• 2003

Recently, many researchers have been interested in biometric systems such as fingerprint, handwriting, key-stroke patterns and human iris. From the viewpoint of reliability and robustness, iris recognition is the most attractive biometric system. Moreover, the iris recognition system is a comfortable biometric system, since the video image of an eye can be taken at a distance. In this paper, we discuss human iris recognition, which is based on accurate iris localization, robust feature extraction, and Neural Network classification. The iris region is accurately localized in the eye image using a multiresolution active snake model. For the feature representation, the localized iris image is decomposed using wavelet transform based on dyadic Haar wavelet. Experimental results show the usefulness of wavelet transform in comparison to conventional Gabor transform. In addition, we present a new method for setting initial weight vectors in competitive learning. The proposed initialization method yields better accuracy than the conventional method.

• Journal of the Semiconductor & Display Technology
• /
• v.19 no.2
• /
• pp.100-104
• /
• 2020

Over the past several years, deep learning has been widely used for feature extraction in image and video for various applications such as object classification and facial recognition. This paper introduces an implantation of embedded Linux system for embossed digits recognition using CNN based deep learning methods. For this purpose, we implemented a coin recognition system based on deep learning with the Keras open source library on Raspberry PI. The performance evaluation has been made with the success rate of coin classification using the images captured with ultra-wide angle camera on Raspberry PI. The simulation result shows 98% of the success rate on average.