From Image Analysis to Computer Vision: An Annotated Bibliography, 1955–1979

Computer Vision for Microscopy Applications

“…Typical approaches to machine learning emphasize optimizing classification in just one particular problem. Because of this, typical implementations of pattern recognition algorithms only allow for a limited set of descriptors (Awate et al, 2006;Boland & Murphy, 2001;Cocosco et al, 2004;Dong & Yang, 2002;Jing & Zhang, 2006;Ranzato et al, 2007;Rosenfeld, 2001;Shen & Bai, 2006;Smeulders et al, 2000). A limited number of features is desirable because it lowers the computational cost, and reduces the dimensionality of the feature space used in classification.…”

“…The feature set can become inapplicable when new images deviate significantly from those the classifier was trained on, or if they are from a different imaging modality. A general computer-vision approach requires an alternative to task-specific or manual feature selection (Rosenfeld, 2001). It should use a large feature set in an application-specific context to automatically pick patterns crucial for the given recognition problem ( Fig.…”

Pattern recognition for biomedical imaging and image-guided diagnosis

“…Image features are quantitative representations of image content [5]. Feature Bank (FB) is a set of algorithms for …”

Visual Alphabets: Video Classification by End Users

“…Several techniques are used to determine the patterns that may be perceived from the image [27,30,36,46,47,48]. With most texture analyzes, textural features are derived from the image, instead of describing arrangements of the individual pixels.…”