• 한국정보통신학회논문지
• /
• 제24권1호
• /
• pp.141-144
• /
• 2020

Convolutional neural networks (CNNs) show high performance in computer vision tasks including object detection, but a lot of weight storage and computation is required. In this paper, a pruning scheme is applied to CNNs for object detection, which can remove much amount of weights with a negligible performance degradation. Contrary to the previous ones, the pruning scheme applied in this paper considers the base accelerator architecture. With the consideration, the pruned CNNs can be efficiently performed on an ASIC or FPGA accelerator. Even with the constrained pruning, the resulting CNN shows a negligible degradation of detection performance, less-than-1% point degradation of mAP on VOD0712 test set. With the proposed scheme, CNNs can be applied to objection dtection efficiently.

• Journal of information and communication convergence engineering
• /
• 제20권3호
• /
• pp.226-233
• /
• 2022

Video surveillance is widely used in security surveillance, military navigation, intelligent transportation, etc. Its main research fields are pattern recognition, computer vision and artificial intelligence. This article uses OpenCV to detect and track vehicles, and monitors by establishing an adaptive model on a stationary background. Compared with traditional vehicle detection, it not only has the advantages of low price, convenient installation and maintenance, and wide monitoring range, but also can be used on the road. The intelligent analysis and processing of the scene image using CAMSHIFT tracking algorithm can collect all kinds of traffic flow parameters (including the number of vehicles in a period of time) and the specific position of vehicles at the same time, so as to solve the vehicle offset. It is reliable in operation and has high practical value.

• 한국정보통신학회논문지
• /
• 제26권2호
• /
• pp.325-328
• /
• 2022

A convolutional neural network(CNN) is widely used in the computer vision tasks, but its computing power requirement needs a design of a special circuit. Most of the computations in a CNN can be implemented efficiently in a digital circuit, but the SoftMax layer has operations unsuitable for circuit implementation, which are exponential and logarithmic functions. This paper proposes a new method to integrate the exponential and logarithmic tables of the conventional circuits into a single table. The proposed structure accesses a look-up table (LUT) only with a few maximum values, and the LUT has the result value directly. Our proposed method significantly reduces the space complexity of the SoftMax layer circuit implementation. But our resulting circuit is comparable to the original baseline with small degradation in precision.

• Journal of Information Processing Systems
• /
• 제18권6호
• /
• pp.794-802
• /
• 2022

The recovery of reasonable depth information from different scenes is a popular topic in the field of computer vision. For generating depth maps with better details, we present an efficacious monocular depth prediction framework with coordinate attention and feature fusion. Specifically, the proposed framework contains attention, multi-scale and feature fusion modules. The attention module improves features based on coordinate attention to enhance the predicted effect, whereas the multi-scale module integrates useful low- and high-level contextual features with higher resolution. Moreover, we developed a feature fusion module to combine the heterogeneous features to generate high-quality depth outputs. We also designed a hybrid loss function that measures prediction errors from the perspective of depth and scale-invariant gradients, which contribute to preserving rich details. We conducted the experiments on public RGBD datasets, and the evaluation results show that the proposed scheme can considerably enhance the accuracy of depth prediction, achieving 0.051 for log10 and 0.992 for δ<1.25³ on the NYUv2 dataset.

• International Journal of Computer Science & Network Security
• /
• 제23권7호
• /
• pp.32-38
• /
• 2023

This study deals with arbitration award via modern technical means; because e-Arbitration is deemed to be one of the most important substitute means for the settlement of disputes arising from electronic transactions. This type of arbitration is characterized by fast settlement of disputes, as well as fast enforcement of awards rendered thereon. The researcher seeks to indicate the content of the award, the conditions for rendering it, and to analyze the legal provisions related to its legal basis in the Saudi Law of Arbitration. This study shows that an arbitration award, rendered via modern technical means has a number of advantages, such as fast settlement, less cost, and keeping pace with modern technology, which is an aim of Saudi Arabia Vision 2030. The study also points out certain problems facing arbitration via technical means; however, the most important of which is the insufficiency of some legal rules associated with traditional arbitration, as contained in the Saudi Law of Arbitrator, which are incompatible with or applicable to an arbitration award which is rendered via modern means.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2023년도 제68차 하계학술대회논문집 31권2호
• /
• pp.267-268
• /
• 2023

본 논문은 안드로이드에서 화면 캡쳐와 OCR을 통한 실시간 번역 애플리케이션 개발을 주제로 한다. 코틀린으로 개발된 애플리케이션은 사용자가 원하는 화면 영역을 캡쳐하여 해당 텍스트를 OCR로 추출하고, 구글 Cloud Vision API와 Cloud Translation API를 활용해 번역한다. 이를 통해 외국어 애플리케이션 사용의 편의성을 향상시키고, 정보의 이해와 공유를 도울 수 있음을 제시한다. 이 기술은 더욱 다양한 분야에서의 활용 가능성을 열어놓고 있다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권11호
• /
• pp.4065-4083
• /
• 2021

Super-resolution can improve the clarity of low-resolution (LR) images, which can increase the accuracy of high-level compute vision tasks. Portable devices have low computing power and storage performance. Large-scale neural network super-resolution methods are not suitable for portable devices. In order to save the computational cost and the number of parameters, Lightweight image processing method can improve the processing speed of portable devices. Therefore, we propose the Enhanced Information Multiple Distillation Network (EIMDN) to adapt lower delay and cost. The EIMDN takes feedback mechanism as the framework and obtains low level features through high level features. Further, we replace the feature extraction convolution operation in Information Multiple Distillation Block (IMDB), with Ghost module, and propose the Enhanced Information Multiple Distillation Block (EIMDB) to reduce the amount of calculation and the number of parameters. Finally, coordinate attention (CA) is used at the end of IMDB and EIMDB to enhance the important information extraction from Spaces and channels. Experimental results show that our proposed can achieve convergence faster with fewer parameters and computation, compared with other lightweight super-resolution methods. Under the condition of higher peak signal-to-noise ratio (PSNR) and higher structural similarity (SSIM), the performance of network reconstruction image texture and target contour is significantly improved.

• 센서학회지
• /
• 제21권4호
• /
• pp.249-255
• /
• 2012

Among the human senses, olfactory information still does not have a proper data presentation method unlike that regarding vision and auditory information. It makes presenting the sense of smell into multimedia information impossible, which may be an exploratory field in human computer interaction. In this paper, we propose an olfactory information presentation method, which is a way to use smell as multimedia information, and show an application for scent generation and odor display using a web service. The olfactory information can present smell characteristics such as intensity, persistence, hedonic tone, and odor description. The structure of data format based on olfactory information can also be organized according to data types such as integer, float, char, string, and bitmap. Furthermore, it can be used for data transmitting via a web service and for odor display using a scent generator. The scent generator, which can display information of smell, is developed to generate 6 odors using 6 aroma solutions and a diluted solution with 14 micro-valves and a micropump. Throughout the experiment, we confirm that the remote user can grasp information of smell transmitted by messenger service and request odor display to the computer controlled scent generator. It contributes to enlarge existing virtual reality and to be proposed as a standard reference method regarding olfactory information presentation for future multimedia technology.

• 반도체디스플레이기술학회지
• /
• 제15권2호
• /
• pp.61-65
• /
• 2016

Facial feature detection is a fundamental function in the field of computer vision such as security, bio-metrics, 3D modeling, and face recognition. There are many algorithms for the function, active shape model is one of the most popular local texture models. This paper addresses issues related to face detection, and implements an efficient extraction algorithm for extracting the facial feature points to use on iOS platform. In this paper, we extend the original ASM algorithm to improve its performance by four modifications. First, to detect a face and to initialize the shape model, we apply a face detection API provided from iOS CoreImage framework. Second, we construct a weighted local structure model for landmarks to utilize the edge points of the face contour. Third, we build a modified model definition and fitting more landmarks than the classical ASM. And last, we extend and build two-dimensional profile model for detecting faces within input images. The proposed algorithm is evaluated on experimental test set containing over 500 face images, and found to successfully extract facial feature points, clearly outperforming the original ASM.

• 반도체디스플레이기술학회지
• /
• 제15권4호
• /
• pp.97-102
• /
• 2016

Image stitching is a process of combining two or more images with overlapping area to create a panorama of input images, which is considered as an active research area in computer vision, especially in the field of augmented reality with 360 degree images. Image stitching techniques can be categorized into two general approaches: direct and feature based techniques. Direct techniques compare all the pixel intensities of the images with each other, while feature based approaches aim to determine a relationship between the images through distinct features extracted from the images. This paper proposes a novel image stitching method based on feature pixels with approximated clustering filter. When the features are extracted from input images, we calculate a meaning of the minutiae, and apply an effective feature extraction algorithm to improve the processing time. With the evaluation of the results, the proposed method is corresponding accurate and effective, compared to the previous approaches.