Visual Perception Quiz

Visual Perception

• This unit covers top-down processing, the Gestalt approach, heuristics vs. algorithms, and evidence that perception is an intelligent activity.

Top-Down Processing: Gestalt Principles

• Perceptual organization occurs for hearing and speech segmentation.
• Gestalt approach applies to speech.
• We group or organize speech into meaningful words based on our knowledge of language.
• When listening to a foreign language, words seem continuous but are distinct to the speaker of the language.
• Examples: "thank you," "terima kasih," "xiexie," "kamsiah," "tochear," "arigatogozaimas," "merci," "gracias denada," "I scream, you scream, we all scream for ice cream."

Heuristics vs. Algorithms

• Gestalt principles provide "best guess" predictions, but may not be accurate all the time.
• In problem-solving, two approaches exist:
  • Heuristics: A best guess answer to a problem (shortcut approach). Fast, not time-consuming, but may not lead to correct solutions every time.
  • Algorithms: A procedure that guarantees solving a problem. Time-consuming; step-by-step procedures.

Evidence That Perception Is an Intelligent Activity

• Perception involves behavioral and physiological evidence.

Main Components in Visual Perception (Human vs. Computer)

• Scenes, image acquisition, image processing, and the brain (for humans) are contrasted to the camera and computer (for computers). 

Why is Perception Hard for Computers?

• Four reasons:
  • The stimulus received from receptors might be ambiguous.
  • Objects need to be separated intelligently.
  • Parts of an object can be hidden.
  • Changes in lightness and darkness can be unclear.
• Humans have built-in intelligence within their perceptual systems.

Where Does Our Perceptual Intelligence Come From?

• Knowledge from experience (memories formed since birth).
• Neurons in the brain respond to faces and other objects in the environment, turning the brain into a processor adapted to each object in the environment.
  • Neurons specialize through evolution and experience to adapt to stimulus environment.

Experience-Dependent Plasticity

• Experiment by Gauthier (1999): fMRI used to measure activity in the FFA (fusiform face area) when subjects viewed human faces vs. "greebles."

  • Results (before training): FFA neurons activate more to human faces and less to greebles.
  • Training: subjects trained to recognize "greebles" over several days.
  • Results (after training): FFA neurons responded almost as strongly to both human faces and greebles.

• Experience and learning allows neurons to specialize and respond to different stimulus, creating experience-dependent plasticity.

Summary

• Top-down processing uses Gestalt principles, heuristics, and algorithms.
• Four factors make visual perception difficult for computers when compared to humans: ambiguous stimuli, intelligent object separation, hidden object parts, and unclear lightness/darkness changes.
• Neurons in the temporal lobe (e.g., FFA) can be trained to specialize in perceiving specific objects, like faces in this scenario providing evidence for visual perception in the brain.

Readings and Assignments

• Read Chapter 3 (pages 55-95) of Goldstein's textbook.
• Create a concept map/mindmap of Chapter 3.
• Complete the COGLAB - Visual Search activity.