Visual Perception: Size and Color Constancy

Fundamentals of Visual Constancy

• Constancy is defined as perceiving an object as having constant physical properties (such as size and color) despite a continuously varying retinal response.
• The retinal response changes dynamically based on environmental variables such as distance and illumination, but human perception of the object's intrinsic properties remains stable.
• Demonstration of Size Variation vs. Perceived Size:
  • Holding a hand directly in front of the face yields a large retinal response ($RR$).
  • Moving the hand farther away yields a small retinal response ($RR$).
  • Bringing the hand back close yields a large retinal response ($RR$).
  • Despite the fluctuating retinal response size (growing big as it moves in, small as it moves away), the hand itself is not perceived as changing size.
• Demonstration of Color Variation vs. Perceived Color:
  • Placing a hand under a specific light source yields a particular retinal response.
  • Shifting the hand under varying light sources produces different retinal responses.
  • Despite the varying retinal response caused by illumination changes, the hand is not perceived as changing color.
• General Rule of Retinal Ambiguity:
  • Retinal response varies directly with distance and illumination.
  • The physical object and human perception of that object remain constant.

Theoretical Model of Retinal Ambiguity

• A given retinal response ($RR$) recorded by an eye is inherently ambiguous regarding the physical identity and scale of the source object.
• A single retinal response $RR$ could correspond to:
  • A hand.
  • A tree.
  • Millions-of-years-old rocks (such as those at CSU).
• Simplified Two-Object Model:
  • Given an ambiguous retinal response $RR$, the visual system must choose between two primary structural possibilities:
  • Object A: An object that is small and close.
  • Object B: An object that is big and far.
• Size Constancy Definition:
  • Size constancy is the process of using an object's distance to determine its actual size.
  • It allows the visual system to maintain a constant perception of size even when the retinal response varies.
  • Key functional dependency: To perceive an object's size, the visual system must first establish its distance.

Depth Cues and Size Constancy

• Linear Convergence:
  • A primary depth cue provided by parallel lines (such as road edges) angling inward as they recede into the distance.
• Three-Car Perceptual Demonstration:
  • In an image containing three cars displayed across converging road lines, all three cars produce the exact same physical retinal response size on the eye/page.
  • Depth cues (such as linear convergence) signal that the top car is farthest away and the bottom car is closest.
  • The visual system treats the bottom car as Object $A$ (close and small) and the top car as Object $B$ (big and far).
  • Because the cars are perceived at different distances, they are perceived as completely different physical sizes despite identical retinal responses.
• Core Principle of Perceived Size:
  • Retinal response alone is insufficient and ambiguous for determining size.
  • Once distance is determined by depth cues, that distance directly dictates perceived size.

The Ponzo Illusion: Structure, Mechanics, and Exam Requirements

• Structural Overview:
  • Features two converging vertical/diagonal lines (representing receding road edges or railroad tracks) and two identical horizontal lines placed between them at different heights.
• Perceptual Effect:
  • The top horizontal line is almost universally perceived as longer/bigger than the bottom horizontal line.
  • Physical reality: Both horizontal lines are identical in length.
• Explanatory Mechanism:
  • The converging outer lines provide the depth cue of linear convergence, signaling distance.
  • The visual system infers that the top line is farther away than the bottom line.
  • Because the top line produces the same retinal response as the bottom line despite being perceived as farther away, the visual system infers that the top line must be physically larger.

Step-by-Step Exam Guide for the Ponzo Illusion

• Drawing Procedure:
  1. Draw the horizontal lines first to ensure they are identical in size.
  2. Draw the outer road lines second, angling them inward toward the top to create linear convergence.
  3. Label the horizontal lines (e.g., line $A$ and line $B$).
• Three-Part Written Explanation Rubric:
  • Part 1: Objective Reality (Truth): Explicitly state that both horizontal lines are the exact same size ($\text{line } A = \text{line } B$).
  • Part 2: Perceptual Experience: Describe what is perceived (the top line appears larger than the bottom line).
  • Part 3: Causal Mechanism (Why): Explain via size constancy. State that to determine size, the visual system must first determine distance. Because the visual system perceives the top line as farther away due to linear convergence, it infers that the top line must be physically larger.
  • Note: Merely naming "size constancy" without explaining distance inference yields only partial credit on short-answer exams.

Empirical Research and Real-World Depth Illusions

• Research by Maz Brokers:
  • Maz Brokers is a scientist at the University of Wisconsin-Madison studying visual perception.
  • Focuses on how the brain processes 2D visual information from both eyes in the visual cortex to map out 3D space, determining object distance and speed.
  • Conducted field experiments at a Los Angeles rail yard to evaluate perspective cues.
• Brain Assumptions and Shortcuts:
  • Parallel lines converge toward a vanishing point on the horizon due to perspective.
  • The visual cortex utilizes familiar line patterns (such as wall-ceiling intersections, doorway frames, and window edges) as cognitive shortcuts to construct spatial maps efficiently.
  • Because the brain relies on these efficient assumptions, it can overcompensate when presented with artificial perspective cues, leading to misperceptions.
  • In real-world environments, these mechanisms are generally highly accurate, preventing people from running into obstacles.
• Real-World Ponzo Experiment on Railroad Tracks:
  • Two physical bars of identical length are placed on actual train tracks.
  • The top bar appears visibly longer because the brain assumes it sits farther down the track.
  • The brain calculates that if an object were truly farther away while maintaining that specific retinal size, it would have to be physically larger, thereby overcompensating to produce a larger perceived size.
• Tabletop Illusion:
  • Involves two tables (left and right) with identical surface dimensions.
  • Vertical perspective lines receding into the background trick the brain into perceiving vertical shapes as significantly thinner and longer than identical horizontal shapes.
• Contextual Note on Academic Ski Conference:
  • Academic conference presentations ran from 4:00 PM to 8:00 PM daily at a ski resort (coinciding with slope closing times at 4:00 PM).
  • Illustrates professional networking dynamics occurring in gondolas versus standard hotel bars.

The Monster Illusion and General Size Constancy Laws

• The Monster Illusion Mechanics:
  • A variation of the Ponzo illusion featuring two identical cartoon monster graphics in a hallway drawn with linear convergence.
  • The monster positioned toward the upper background appears significantly larger than the monster in the foreground.
  • Taking a screenshot of the monster and moving it from the back to the front visually demonstrates that the physical graphic size is identical, even though depth cues make it appear to shrink as it moves forward.
• Universal Law of Size Constancy:
  • If two objects produce the exact same retinal response size ($RR$), but one object is perceived as farther away, the farther object will be perceived as larger.
  • Formulaic relationship: $\text{Perceived Size} = f(\text{Retinal Response}, \text{Perceived Distance})$.
  • Constancy operates continuously in real time for all visual processing, not just during visual illusions.
  • Illusions do not constitute size constancy itself; rather, illusions are clever experimental tools designed to demonstrate the internal mechanics of size constancy.

Fundamentals of Color Constancy

• Definition of Color Constancy:
  • Color constancy is using an object's illumination to perceive its actual color.
  • Illumination: The light reflecting off an object.
• Functional Mechanism:
  • The visual system measures total light reflection, takes into account environmental illumination, and subtracts out the illumination to reveal the object's true reflectance/color.
  • Without color constancy, an object would be perceived as a completely new object every time ambient lighting shifted.
• Photoreceptor Firing Rate Grid Analysis:
  • Image section analysis showing a gray cup on a gray table against a black background.
  • Photoreceptor firing rates yield identical intensity values of $4$ for two distinct regions:
  1. The black background under bright illumination.
  2. A section of the gray cup located in deep shadow.
  • Identical retinal response values ($RR = 4$) present two computational interpretations for the visual system:
  • Option A: A dark object under bright illumination.
  • Option B: A bright object under dark illumination/shadow.
  • The visual system subtracts the bright illumination from region 1 to perceive black, and subtracts the shadow illumination from region 2 to perceive gray.

Case Study: The Dress Illusion ("Blue-Black" vs. "White-Gold")

• Perceptual Split:
  • Viewers looking at the exact same image perceive either a blue and black dress or a white and gold dress.
  • Blue/Black Viewers: Visual system infers warm/yellow artificial illumination and subtracts yellow light, perceiving a blue and black dress.
  • White/Gold Viewers: Visual system infers cool/blue natural illumination (such as light coming through a window) and subtracts blue light, perceiving a white and gold dress.
  • Switching between perceptions is possible for a small minority of viewers, but rare.
• Ambiguity of Illumination:
  • Environmental illumination in the photograph is ambiguous.
  • Illumination inference is executed as an unconscious inference by the visual cortex.
  • When the background lighting context is artificially clarified or cropped out, all viewers perceive the identical baseline color.
• Refuted Explanations for the Dress Illusion:
  1. Screen/Monitor Differences: Incorrect. Individuals looking at the exact same monitor at the exact same time still experience different color perceptions.
  2. Photoreceptor Ratios: Incorrect. Although human retinas possess varying ratios of red, green, and blue cone photoreceptors, these ratio variations do not correlate with whether a viewer sees blue/black or white/gold.
• Scientific Consensus:
  • The illusion is caused entirely by color constancy mechanisms processing an ambiguous illumination context.

Further Color Constancy Illusions

• Rubik's Cube Tile Illusion:
  • A 3D cube features tiles on a top face under bright light and tiles on a side face under dark shadow.
  • The center tile on the top face is perceived as brown.
  • The center tile on the shadowed side face is perceived as bright orange.
  • Physical Reality: Both central tiles reflect the exact same wavelength and intensity of light to the retina.
  • Perceptual Mechanism: The visual system accounts for bright ambient light on the top versus shadow on the side, subtracting the shadow to render the side tile as bright orange.

Synthesis and Exam Review

• System Requirement for Constancy:
  • Constancy is necessary because retinal responses are constantly changing due to object distance shifts, body movement, and illumination changes.
  • Constancy allows humans to perceive stable, constant object properties (size, color, shape).
  • In a simulated environment completely devoid of distance cues and illumination cues, visual constancy fails completely, making sense-making impossible.
• Sample Exam Question Formats:
  • Direct Conceptual Question: "To perceive an object's color, what must the visual system first detect or subtract out?" (Answer: Environmental illumination / light).
  • Applied Short-Answer Question: "Describe and diagram an example that demonstrates size constancy (e.g., the Ponzo illusion), including the truth, the perception, and the visual mechanism."