KOSEN 한인과학기술자네트워크

IMAGE COMPUTING

From Computer Vision and Image Processing: A Practical Approach Using CVIPtools,
S.E. Umbaugh, Prentice Hall, 1998

컴퓨터를 사용하여 시각 정보인 영상을 획득하고 처리하는 분야를 총칭
Image computing의 두 가지 중요한 범주: Computer Vision, Image Processing

Computer Vision
디지털 영상을 분석하거나 특징 정보(feature)를 추출함으로써 영상의 내용을 분석하고 이해하는 분야
분석된 영상 내용은 컴퓨터를 기반으로하는 비젼 시스템이 활용

▩ Field of Computer Vision
  ▷ Feature Extraction
      Process of acquiring higher level image information (shape,color)
  ▷ Pattern Recognition
      Act of taking this higher-level information and identifying objects within the image

▩ Applications of Computer Vision
  - factory automation system
  - medical diagnostic system
  - automatic identification - DNA analysis, fingerprints identification, retinal identification
  - weather prediction & intelligent transport system

Digital Image Processing
입력 영상으로부터 인간 시각의 관점에서 보다 향상된 양질의 영상을 생성하는 분야
Computer Vision과는 달리 처리된 결과 영상의 인간의 시각 관점에서 활용

▩ Fields of image processing
  ▷ Image restoration
  ▷ Image enhancement
  ▷ Image compression

▩ Applications of image processing
  ▷ medical - PET, CT, MRI
  ▷ biological research
  ▷ entertainment industry, etc

From Image Database: Search and Retrieval of Digital Imagery,
Edited by Vittorio Castelli and Lawrence D. Bergman

Two related fields of Image Retrieval

▩ Database management : text-based
  ⊙ Image were annotated by key words and stored as retrieval keys in traditional database
  ⊙ Problems - when the size of image database becomes very large
     ▷ Amount of labor involved in image annotation
     ▷ Difficulty of capturing the rich content of images using a small number of key words
     ▷ The subjectivity of human perception

▩ Computer vision : visual properties of the data

Content-Based Image Retrieval (CBIR)
  ⊙ “to find all images in a given database depicting scenes or objects of some specified type by users”
  ⊙ automatic indexing by summarizing their low-level visual contents like color, texture and shape
  ⊙ Low-level features, extracted by a computer, are substituted for higher-level, text-based manual annotations or key words
  ⊙ Images are searched and matched based on the low-level features
  ⊙ Ideally CBIR system should automatically extract and index the semantic content of images
  ⊙ Automatic object recognition and classification are still among the most difficult problems in computer vision
  ⊙ Thus low-level features like color, texture and shape, are widely used for CBIR

멀티미디어컨텐츠

From 멀티미디어 배움터 - 최윤철, 고견, 임순범 공저, 생능출판사