• Title/Summary/Keyword: scene detection

Search Result 519, Processing Time 0.027 seconds

Light weight architecture for acoustic scene classification (음향 장면 분류를 위한 경량화 모형 연구)

  • Lim, Soyoung;Kwak, Il-Youp
    • The Korean Journal of Applied Statistics
    • /
    • v.34 no.6
    • /
    • pp.979-993
    • /
    • 2021
  • Acoustic scene classification (ASC) categorizes an audio file based on the environment in which it has been recorded. This has long been studied in the detection and classification of acoustic scenes and events (DCASE). In this study, we considered the problem that ASC faces in real-world applications that the model used should have low-complexity. We compared several models that apply light-weight techniques. First, a base CNN model was proposed using log mel-spectrogram, deltas, and delta-deltas features. Second, depthwise separable convolution, linear bottleneck inverted residual block was applied to the convolutional layer, and Quantization was applied to the models to develop a low-complexity model. The model considering low-complexity was similar or slightly inferior to the performance of the base model, but the model size was significantly reduced from 503 KB to 42.76 KB.

Target Recognition Algorithm Based on a Scanned Image on a Millimeter-Wave(Ka-Band) Multi-Mode Seeker (스캔 영상 기반의 밀리미터파(Ka 밴드) 복합모드 탐색기 표적인식 알고리즘 연구)

  • Roh, Kyung A;Jung, Jun Young;Song, Sung Chan
    • The Journal of Korean Institute of Electromagnetic Engineering and Science
    • /
    • v.30 no.2
    • /
    • pp.177-180
    • /
    • 2019
  • To improve the accuracy rate of guided weapons, many studies have been conducted on the accurate detection and identification of targets from sea clutter. Because of the variety and complicated characteristics of both sea-clutter and target signals, an active target recognition technique is required. In this study, we propose an algorithm to distinguish clutter and recognize targets by applying a fractal signature(FS) classifier, which is a fractal dimension, and a high-resolution target image(HRTI) classifier, which applies scene matching to an image formed from a scanned image. Simulation results using the algorithm revealed that the HRTI classifier recognized targets 1 and 2 at a 100 % rate, whereas the FS classifier recognized targets 1 and 2 at rates of 90 % and 93 %, respectively.

Text Extraction from Complex Natural Images

  • Kumar, Manoj;Lee, Guee-Sang
    • International Journal of Contents
    • /
    • v.6 no.2
    • /
    • pp.1-5
    • /
    • 2010
  • The rapid growth in communication technology has led to the development of effective ways of sharing ideas and information in the form of speech and images. Understanding this information has become an important research issue and drawn the attention of many researchers. Text in a digital image contains much important information regarding the scene. Detecting and extracting this text is a difficult task and has many challenging issues. The main challenges in extracting text from natural scene images are the variation in the font size, alignment of text, font colors, illumination changes, and reflections in the images. In this paper, we propose a connected component based method to automatically detect the text region in natural images. Since text regions in mages contain mostly repetitions of vertical strokes, we try to find a pattern of closely packed vertical edges. Once the group of edges is found, the neighboring vertical edges are connected to each other. Connected regions whose geometric features lie outside of the valid specifications are considered as outliers and eliminated. The proposed method is more effective than the existing methods for slanted or curved characters. The experimental results are given for the validation of our approach.

Fast Scene Understanding in Urban Environments for an Autonomous Vehicle equipped with 2D Laser Scanners (무인 자동차의 2차원 레이저 거리 센서를 이용한 도시 환경에서의 빠른 주변 환경 인식 방법)

  • Ahn, Seung-Uk;Choe, Yun-Geun;Chung, Myung-Jin
    • The Journal of Korea Robotics Society
    • /
    • v.7 no.2
    • /
    • pp.92-100
    • /
    • 2012
  • A map of complex environment can be generated using a robot carrying sensors. However, representation of environments directly using the integration of sensor data tells only spatial existence. In order to execute high-level applications, robots need semantic knowledge of the environments. This research investigates the design of a system for recognizing objects in 3D point clouds of urban environments. The proposed system is decomposed into five steps: sequential LIDAR scan, point classification, ground detection and elimination, segmentation, and object classification. This method could classify the various objects in urban environment, such as cars, trees, buildings, posts, etc. The simple methods minimizing time-consuming process are developed to guarantee real-time performance and to perform data classification on-the-fly as data is being acquired. To evaluate performance of the proposed methods, computation time and recognition rate are analyzed. Experimental results demonstrate that the proposed algorithm has efficiency in fast understanding the semantic knowledge of a dynamic urban environment.

Scene-based Nonuniformity Correction for Neural Network Complemented by Reducing Lense Vignetting Effect and Adaptive Learning rate

  • No, Gun-hyo;Hong, Yong-hee;Park, Jin-ho;Jhee, Ho-jin
    • Journal of the Korea Society of Computer and Information
    • /
    • v.23 no.7
    • /
    • pp.81-90
    • /
    • 2018
  • In this paper, reducing lense Vignetting effect and adaptive learning rate method are proposed to complement Scribner's neural network for nuc algorithm which is the effective algorithm in statistic SBNUC algorithm. Proposed reducing vignetting effect method is updated weight and bias each differently using different cost function. Proposed adaptive learning rate for updating weight and bias is using sobel edge detection method, which has good result for boundary condition of image. The ordinary statistic SBNUC algorithm has problem to compensate lense vignetting effect, because statistic algorithm is updated weight and bias by using gradient descent method, so it should not be effective for global weight problem same like, lense vignetting effect. We employ the proposed methods to Scribner's neural network method(NNM) and Torres's reducing ghosting correction for neural network nuc algorithm(improved NNM), and apply it to real-infrared detector image stream. The result of proposed algorithm shows that it has 10dB higher PSNR and 1.5 times faster convergence speed then the improved NNM Algorithm.

Body fluid identification in forensics

  • An, Ja-Hyun;Shin, Kyoung-Jin;Yang, Woo-Ick;Lee, Hwan-Young
    • BMB Reports
    • /
    • v.45 no.10
    • /
    • pp.545-553
    • /
    • 2012
  • Determination of the type and origin of the body fluids found at a crime scene can give important insights into crime scene reconstruction by supporting a link between sample donors and actual criminal acts. For more than a century, numerous types of body fluid identification methods have been developed, such as chemical tests, immunological tests, protein catalytic activity tests, spectroscopic methods and microscopy. However, these conventional body fluid identification methods are mostly presumptive, and are carried out for only one body fluid at a time. Therefore, the use of a molecular genetics-based approach using RNA profiling or DNA methylation detection has been recently proposed to supplant conventional body fluid identification methods. Several RNA markers and tDMRs (tissue-specific differentially methylated regions) which are specific to forensically relevant body fluids have been identified, and their specificities and sensitivities have been tested using various samples. In this review, we provide an overview of the present knowledge and the most recent developments in forensic body fluid identification and discuss its possible practical application to forensic casework.

Comparisons of Object Recognition Performance with 3D Photon Counting & Gray Scale Images

  • Lee, Chung-Ghiu;Moon, In-Kyu
    • Journal of the Optical Society of Korea
    • /
    • v.14 no.4
    • /
    • pp.388-394
    • /
    • 2010
  • In this paper the object recognition performance of a photon counting integral imaging system is quantitatively compared with that of a conventional gray scale imaging system. For 3D imaging of objects with a small number of photons, the elemental image set of a 3D scene is obtained using the integral imaging set up. We assume that the elemental image detection follows a Poisson distribution. Computational geometrical ray back propagation algorithm and parametric maximum likelihood estimator are applied to the photon counting elemental image set in order to reconstruct the original 3D scene. To evaluate the photon counting object recognition performance, the normalized correlation peaks between the reconstructed 3D scenes are calculated for the varied and fixed total number of photons in the reconstructed sectional image changing the total number of image channels in the integral imaging system. It is quantitatively illustrated that the recognition performance of the photon counting integral imaging system can be similar to that of a conventional gray scale imaging system as the number of image viewing channels in the photon counting integral imaging (PCII) system is increased up to the threshold point. Also, we present experiments to find the threshold point on the total number of image channels in the PCII system which can guarantee a comparable recognition performance with a gray scale imaging system. To the best of our knowledge, this is the first report on comparisons of object recognition performance with 3D photon counting & gray scale images.

A study on characteristics of latent fingerprint detection on vinyl leather (합성가죽에서의 잠재지문 현출)

  • Choi, Yong-Bok;Park, Sung-Woo
    • Analytical Science and Technology
    • /
    • v.24 no.1
    • /
    • pp.45-50
    • /
    • 2011
  • While the crime scene investigation, evidences are found to search with the naked eyes and to use science equipments. Crime evidences, which are used as judgement, have various things: blood, sperm, saliva, hair, fingerprint, fiber, soil, paint, and so on, especially, fingerprints could take through crime scene investigation, and in there, it could distinguish individually. In this case study, sort of leathers: a wallet, a belt, and a sofa, which are needed special identification even thought people experience easily in usual day, are tried many of the identification tests. As a result, using fluorescent magnetic powder method or fluorescent powder method as identification of many leathers is most efficient.

Indexing Considering Video Rating of Scenes in Video (동영상의 장면별 비디오 등급을 고려한 색인)

  • Kim Young-Bong
    • Journal of Game and Entertainment
    • /
    • v.2 no.2
    • /
    • pp.51-60
    • /
    • 2006
  • Recently, many streaming videos including drama, music videos, and movies have been extensively given on the web. Such video services are on negative lines in any service restriction depending on the age of users and then whole part of a video have been restricted considering the age of users. Therefore, in this paper, we will present a new method that provides the access depending on the ages of users and also sets the video rating of each scene in a video. To get this restricted access for video, we will first divide a streaming video into many scenes using histogram techniques. Each scene gets an access control depending on the nudity level. Finally, we will make the video indexing including the access level depending on its nudity level and then hide restricted scenes using several masks in playing that streaming video.

  • PDF

Improved Similarity Detection Algorithm of the Video Scene (개선된 비디오 장면 유사도 검출 알고리즘)

  • Yu, Ju-Won;Kim, Jong-Weon;Choi, Jong-Uk;Bae, Kyoung-Yul
    • The Journal of the Korea Contents Association
    • /
    • v.9 no.2
    • /
    • pp.43-50
    • /
    • 2009
  • We proposed similarity detection method of the video frame data that extracts the feature data of own video frame and creates the 1-D signal in this paper. We get the similar frame boundary and make the representative frames within the frame boundary to extract the similarity extraction between video. Representative frames make blurring frames and extract the feature data using DOG values. Finally, we convert the feature data into the 1-D signal and compare the contents similarity. The experimental results show that the proposed algorithm get over 0.9 similarity value against noise addition, rotation change, size change, frame delete, frame cutting.