• ETRI Journal
• /
• v.33 no.4
• /
• pp.569-579
• /
• 2011

This paper presents a novel approach for human activity recognition (HAR) using the joint angles from a 3D model of a human body. Unlike conventional approaches in which the joint angles are computed from inverse kinematic analysis of the optical marker positions captured with multiple cameras, our approach utilizes the body joint angles estimated directly from time-series activity images acquired with a single stereo camera by co-registering a 3D body model to the stereo information. The estimated joint-angle features are then mapped into codewords to generate discrete symbols for a hidden Markov model (HMM) of each activity. With these symbols, each activity is trained through the HMM, and later, all the trained HMMs are used for activity recognition. The performance of our joint-angle-based HAR has been compared to that of a conventional binary and depth silhouette-based HAR, producing significantly better results in the recognition rate, especially for the activities that are not discernible with the conventional approaches.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.5
• /
• pp.187-194
• /
• 2021

A person's behavioral recognition is the recognition of what a person does according to joint movements. To this end, we utilize computer vision tasks that are utilized in image processing. Human behavior recognition is a safety accident response service that combines deep learning and CCTV, and can be applied within the safety management site. Existing studies are relatively lacking in behavioral recognition studies through human joint keypoint extraction by utilizing deep learning. There were also problems that were difficult to manage workers continuously and systematically at safety management sites. In this paper, to address these problems, we propose a method to recognize risk behavior using only joint keypoints and joint motion information. AlphaPose, one of the pose estimation methods, was used to extract joint keypoints in the body part. The extracted joint keypoints were sequentially entered into the Long Short-Term Memory (LSTM) model to be learned with continuous data. After checking the behavioral recognition accuracy, it was confirmed that the accuracy of the "Lying Down" behavioral recognition results was high.

• Journal of the Optical Society of Korea
• /
• v.13 no.4
• /
• pp.445-450
• /
• 2009

Fingerprint recognition using a joint transform correlator (JTC) is the most well-known technology among optical fingerprint recognition methods. The JTC method optically compares the reference fingerprint image with the sample fingerprint image then examines match or non-match by acquiring a correlation peak. In contrast to the JTC method, this paper presents a new method to examine fingerprint recognition by producing a computer generated hologram (CGH) of those two fingerprint images and directly comparing them. As a result, we present some parameters to show that fingerprint recognition capability of the CGH direct comparison method is superior to that of the JTC method.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1662-1665
• /
• 2004

A real-time joint transform correlator by using JPEG-compressed reference images is proposed as practical solution to storage problem and improvement of processing time of automatic target recognition system [1]. Effects of compression on recognition performance of join transform correlator are quantitatively investigated under situations where the target is suffered from noise and has contrast difference with respect to the reference. Two images with different spatial-frequency contents and contrast were used as the test scenes. The simulation results show that, the recognition performance of joint transform correlator by using the compressed reference images with high spatial-frequency components is more sensitive to noise and contrast difference than the low spatial-frequency image.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.6
• /
• pp.2767-2780
• /
• 2016

Video-based human-activity recognition has become increasingly popular due to the prominent corresponding applications in a variety of fields such as computer vision, image processing, smart-home healthcare, and human-computer interactions. The essential goals of a video-based activity-recognition system include the provision of behavior-based information to enable functionality that proactively assists a person with his/her tasks. The target of this work is the development of a novel approach for human-activity recognition, whereby human-body-joint features that are extracted from depth videos are used. From silhouette images taken at every depth, the direction and magnitude features are first obtained from each connected body-joint pair so that they can be augmented later with motion direction, as well as with the magnitude features of each joint in the next frame. A generalized discriminant analysis (GDA) is applied to make the spatiotemporal features more robust, followed by the feeding of the time-sequence features into a Hidden Markov Model (HMM) for the training of each activity. Lastly, all of the trained-activity HMMs are used for depth-video activity recognition.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.2
• /
• pp.1118-1133
• /
• 2017

Computer vision-based human activity recognition (HAR) has become very famous these days due to its applications in various fields such as smart home healthcare for elderly people. A video-based activity recognition system basically has many goals such as to react based on people's behavior that allows the systems to proactively assist them with their tasks. A novel approach is proposed in this work for depth video based human activity recognition using joint-based motion features of depth body shapes and Deep Belief Network (DBN). From depth video, different body parts of human activities are segmented first by means of a trained random forest. The motion features representing the magnitude and direction of each joint in next frame are extracted. Finally, the features are applied for training a DBN to be used for recognition later. The proposed HAR approach showed superior performance over conventional approaches on private and public datasets, indicating a prominent approach for practical applications in smartly controlled environments.

• Journal of the Optical Society of Korea
• /
• v.15 no.2
• /
• pp.140-145
• /
• 2011

Joint transform correlator (JTC) is a well known tool for color pattern recognition for a color image. Color images have red, green and blue components, thus in conventional JTC, three input channels of these color components are necessary for color pattern recognition. This paper proposes a new technique of color pattern recognition by decomposing the color image into three color components and recombining those components into a single gray image in the input plane. This new technique needs single input channel and single output CCD camera, thus a simple JTC can be used. We present various kinds of simulated results to show that our newly proposed technique can accurately recognize and discriminate color differences.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.413-415
• /
• 2022

Since human joints can be used as useful information for analyzing human behavior as a component of the human body, many studies have been conducted on human action recognition using joint information. However, it is a very complex problem to recognize human action that changes every moment using only each independent joint information. Therefore, an additional information extraction method to be used for learning and an algorithm that considers the current state based on the past state are needed. In this paper, we propose a human action recognition technique considering the positional relationship of connected joints and the change of the position of each joint over time. Using the pre-trained joint extraction model, position information of each joint is obtained, and bone information is extracted using the difference vector between the connected joints. In addition, a simplified neural network is constructed according to the two types of inputs, and spatio-temporal features are extracted by adding LSTM. As a result of the experiment using a dataset consisting of 9 behaviors, it was confirmed that when the action recognition accuracy was measured considering the temporal and spatial relationship features of each joint, it showed superior performance compared to the result using only single joint information.

• Proceedings of the KIEE Conference
• /
• 2007.04a
• /
• pp.219-221
• /
• 2007

Recent changes to ubiquitous computing requires more natural human-computer(HCI) interfaces that provide high information accessibility. Hand-gesture, i.e., gestures performed by one 'or two hands, is emerging as a viable technology to complement or replace conventional HCI technology. This paper deals with hand-posture recognition. Hand-posture database construction is important in hand-posture recognition. Human hand is composed of 27 bones and the movement of each joint is modeled by 23 degrees of freedom. Even for the same hand-posture,. grabbed images may differ depending on user's characteristic and relative position between the hand and cameras. To solve the difficulty in defining hand-postures and construct database effective in size, we present a method using a 3D hand model. Hand joint angles for each hand-posture and corresponding silhouette images from many viewpoints by projecting the model into image planes are used to construct the ?database. The proposed method does not require additional equations to define movement constraints of each joint. Also using the method, it is easy to get images of one hand-posture from many vi.ewpoints and distances. Hence it is possible to construct database more precisely and concretely. The validity of the method is evaluated by applying it to the hand-posture recognition system.

• Journal of the Optical Society of Korea
• /
• v.15 no.4
• /
• pp.335-339
• /
• 2011

Recently, we reported a single input channel joint transform correlator for the color pattern recognition which decomposes the input color image into three R, G, and B gray components and adds those components into a single gray image in the input plane. This technique has the merit of a single input channel instead of three input channels. However, we found this technique has some problems with discrimination impossibility in the case of a simple primary color pattern which results in the same gray level through the addition process. Thus, we propose a modified coding technique which selectively recombines the decomposed three R, G, and B gray components instead of the simple adding process. Simulated results show that the modified coding technique can accurately discriminate a variety of kinds of color images.