David Marr
British neuroscientist whose levels of analysis framework structures vision research.
Quotes by David Marr
Trying to understand perception by studying only neurons is like trying to understand bird flight by studying only feathers: It just cannot be done. In order to understand bird flight, we have to understand aerodynamics; only then do the structure of feathers and the different shapes of birds' wings make sense.
A computational theory is an abstract formulation of the problem that the system is trying to solve, and why it is trying to solve it.
The study of information processing systems is not the study of hardware or software, but the study of the problems that these systems solve.
The central problem of vision is to compute from an image of an object a representation of its shape.
The main point is that the brain is an information processing device, and we need to understand the computations it performs.
The three levels of analysis are: computational theory, algorithmic representation, and hardware implementation.
The primary goal of computational vision is to understand how the visual system works.
Vision is the process of discovering from images what is out there in the world, and where it is.
The visual system is not a camera; it is a sophisticated information processing device.
Understanding a complex system requires understanding it at multiple levels of abstraction.
The computational level is the most important level of analysis, because it defines the problem to be solved.
The brain is a machine, and we can understand how it works by studying its computations.
The goal of vision is to derive a useful description of the world from images.
The visual system is an elegant solution to a difficult computational problem.
The study of vision is the study of how the brain creates a representation of the world.
The primary goal of neuroscience should be to understand the computations performed by the brain.
The brain is not just a collection of neurons; it is a computational device.
To understand vision, we must understand the nature of the information that is processed.
The computational approach to vision provides a framework for understanding the visual system.
The visual system is a highly organized and structured system.