• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.53 no.3
• /
• pp.182-187
• /
• 2004

In this paper, we propose a PC based distributed controller for a two axis convey table using real-time micro-kernel. PC, Windows program, gives an easy way to implement wealthy GUI and micro-kernel, ${\mu}$C/OS-II, provides a real-time capability to control devices. We built a real-time distributed control system using ${\mu}$C/OS-II kernel which needs to process the tasks for two motors within the desired time to synchronize the motion. We used both semaphore and message mail box for synchronization. Unlike the previous study where we used step motors for the actuator of two axes convey table, we rebuilt the convey table with DC motors and the dedicated position servo which had built in out lab, and then we implemented a realtime distributed control system by putting the micro-kernel into between PC and position servo. Moreover we developed the PC based graphic user interfaces for generating planar drawing image control. Experimental results also presented to show the Proposed control system is useful.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.6
• /
• pp.21-28
• /
• 2019

Recent years with the development of artificial intelligence and the success of the deep model, they have been deployed in all fields of computer vision. Action recognition, as an important branch of human perception and computer vision system research, has attracted more and more attention. Action recognition is a challenging task due to the special complexity of human movement, the same movement may exist between multiple individuals. The human action exists as a continuous image frame in the video, so action recognition requires more computational power than processing static images. And the simple use of the CNN network cannot achieve the desired results. Recently, the attention model has achieved good results in computer vision and natural language processing. In particular, for video action classification, after adding the attention model, it is more effective to focus on motion features and improve performance. It intuitively explains which part the model attends to when making a particular decision, which is very helpful in real applications. In this paper, we proposed a 3D dense convolutional network based on attention mechanism(ADD-Net), recognition of human motion behavior in the video.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.488-489
• /
• 2019

We propose a method to identify vehicle type by the formula of distance between feature points of vehicle and proportional rate of size. Car images are converted from the basic RGB model to the gray color model. Perform Canny Edge Direction to remove the background image of the car. The desired feature points are obtained through contour extraction.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.686-687
• /
• 2022

Existing clothing product search is possible only with text search, so it is not possible to properly search for the product that the user is looking for, and it takes a lot of time to search for the desired product. In addition, it was difficult to know where to sell unbranded products such as Soho Mall products, and it was difficult to search for detailed attributes of fashion products. To solve this problem, we would like to propose a deep learning technique that can search for the exact information that the user wants through the learned image search.

• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.2
• /
• pp.133-143
• /
• 2023

In this study, we propose a big data set of strawberry pests collected directly for diagnosis model learning and an automatic pest diagnosis model architecture based on deep learning. First, a big data set related to strawberry pests, which did not exist anywhere before, was directly collected from the web. A total of more than 12,000 image data was directly collected and classified, and this data was used to train a deep learning model. Second, the deep-learning-based automatic pest diagnosis module is a module that classifies what kind of pest or disease corresponds to when a user inputs a desired picture. In particular, we propose a model architecture that can optimally classify pests based on a convolutional neural network among deep learning models. Through this, farmers can easily identify diseases and pests without professional knowledge, and can respond quickly accordingly.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.38 no.8
• /
• pp.629-639
• /
• 1989

One of the major issues of the industrial vision systems is lengthy processing time due to bulky data for image coding. To reduce the processing time, a modified chain coding algorithm is proposed in such a way that it is more suitable for hardware implementation. A hardware chain coder is developed and used for learn and recognizing objects by extracting several features. It is shown that the desired vision system is much faster than a typical software based system so that it may be applicable to real-time industrial operations.

• ETRI Journal
• /
• v.45 no.5
• /
• pp.811-821
• /
• 2023

We propose an adaptive unmanned aerial vehicle (UAV)-assisted object recognition algorithm for urban surveillance scenarios. For UAV-assisted surveillance, UAVs are equipped with learning-based object recognition models and can collect surveillance image data. However, owing to the limitations of UAVs regarding power and computational resources, adaptive control must be performed accordingly. Therefore, we introduce a self-adaptive control strategy to maximize the time-averaged recognition performance subject to stability through a formulation based on Lyapunov optimization. Results from performance evaluations on real-world data demonstrate that the proposed algorithm achieves the desired performance improvements.

• Archives of design research
• /
• v.16 no.2
• /
• pp.405-414
• /
• 2003

The modeling analysis for objects placed in a given space can be described objectively when their visual image structure is grasped. It can't be answered without first analyzing the basic program, visual expression. And when the whole aspect of the visual image of the desired interior utensils is presented, the mindset of its designer can be deduced from that. Therefore, the study was based on the lemon juice squeezer, one of the interior kitchen utensils that Philippe Starck designed for Italy ALESSI company. For the study method, putting'The Elements of Dynamic Symmetry' by Prof. Jay Hambidge into practice, 'paradigm' analysis containing the whole'lemon juice squeezer'image was attempted. And to describe it, the visual mark description method by Prof. Bok-Young Kim was used. In conclusion, henceforth, the relationship between interior space and articles, the relationship between object and user, the modeling critique or analysis of the production itself shoud not be intended to be emotional. On the contrary, the study presented an art analyic methodology that can analyze and describe the visual image structure numerically, and confirm the relationship between form and content.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.8
• /
• pp.1-11
• /
• 1998

One of the desired features for the realizations of high quality Information and Telecommunication services in future is "the Sensation of Reality". This will be achieved only with the visual communication based on the 3- dimensional (3-D) moving images. The main difficulties in realizing 3-D moving image communication are that there is no developed data transmission technology for the hugh amount of data involved in 3-D images and no established technologies for 3-D image recording and displaying in real time. The currently known stereoscopic imaging technologies can only present depth, no moving parallax, so they are not effective in creating the sensation of the reality without taking eye glasses. The more effective 3-D imaging technologies for achieving the sensation of reality are those based on the multiview 3-D images which provides the object image changes as the eyes move to different directions. In this paper, a multiview 3-D imaging system composed of 8 CCD cameras in a case, a RGB(Red, Green, Blue) beam projector, and a holographic screen is introduced. In this system, the 8 view images are recorded by the 8 CCD cameras and the images are transmitted to the beam projector in sequence by a signal converter. This signal converter converts each camera signal into 3 different color signals, i.e., RGB signals, combines each color signal from the 8 cameras into a serial signal train by multiplexing and drives the corresponding color channel of the beam projector to 480Hz frame rate. The beam projector projects images to the holographic screen through a LCD shutter. The LCD shutter consists of 8 LCD strips. The image of each LCD strip, created by the holographic screen, forms as sub-viewing zone. Since the ON period and sequence of the LCD strips are synchronized with those of the camera image sampling adn the beam projector image projection, the multiview 3-D moving images are viewed at the viewing zone.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.4C
• /
• pp.253-261
• /
• 2005

In this paper, we propose an efficient block-based disparity estimation algorithm fur multiple view image coding in EE2 and EE3 in 3DAV. The proposed method emphasizes on visual quality improvement to satisfy the requirements for multiple view generation. Therefore, we perform an adaptive disparity estimation that constructs variable blocks by considering given image features. Examining neighboring features around desired block search range is set up to decrease complexity and additional information than only using quad-tree coding through applying binary-tree and quad-tree coding by taking into account stereo image feature having big disparity. The experimental results show that the proposed method improves PSNR about 1 to 2dB compared to existing other methods and decreases computational complexity up to maximum 68 percentages than FBMA.