Computer Vision

Livro digital

Título:
Computer Vision

Autor:
Dana Ballard, Chris Brown

Categoria:
Tecnologia > IA

Doador:
Raffaello D. N.

Sinopse:
Computer Vision asks a core question: how can an image become a meaningful description of the world? Its table of contents makes that challenge concrete, moving from image formation and early processing into boundary detection, region growing, texture, and motion before advancing to geometric and relational structures. The book then pushes past low-level vision into representation and reasoning. It covers edge operators, stereo, photometric stereo, optical flow, the Hough method, split-and-merge region methods, texture grammars, 2D and 3D shape representations, semantic nets, matching, inference, and goal achievement, showing how visual data and prior knowledge interact across the pipeline. That range makes it especially valuable as a classic academic text rather than a narrow techniques manual. Readers get both the physics-and-algorithms side of vision and the AI side of interpretation, with enough breadth to understand how early computer vision systems were organized and why representation mattered so much to the field.