Related Experiment Video
Updated: Dec 10, 2025
07:11
Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
Published on: December 8, 2023
2.2K
What do adversarial images tell us about human vision?
Marin Dujmović1, Gaurav Malhotra1, Jeffrey S Bowers1
1School of Psychological Science, University of Bristol, Bristol, United Kingdom.
Elife
|September 3, 2020
Summary
Deep convolutional neural networks (DCNNs) show weak agreement with human vision when interpreting adversarial images. This challenges their use as models of human perception, suggesting further research is needed.
Area of Science:
- Cognitive Science
- Computer Vision
- Neuroscience
Background:
- Deep convolutional neural networks (DCNNs) are often considered models of human and primate vision.
- Adversarial images, which are imperceptible to humans but fool DCNNs, present a challenge to this view.
- Previous research suggested potential similarities in how humans and DCNNs interpret these images.
Purpose of the Study:
- To re-evaluate the agreement between human perception and DCNNs when interpreting adversarial images.
- To investigate the influence of image generation methods and experimental design on human-DCNN agreement.
- To determine if adversarial images remain a challenge for DCNNs as models of human vision.
Main Methods:
- Reanalysis of existing data from a prominent study.
- Conducting five new experiments with controlled image generation and selection.
- Systematic evaluation of human-DCNN agreement across different adversarial image types.
Main Results:
- Human-DCNN agreement is significantly weaker and more variable than previously reported.
- The observed agreement is highly dependent on the specific adversarial images used and the experimental setup.
- Certain adversarial image generation methods result in no discernible agreement between humans and DCNNs.
Conclusions:
- Adversarial images continue to pose a significant challenge to the use of DCNNs as accurate models of human vision.
- The findings question the extent to which current DCNNs capture the nuances of human visual processing.
- Further research is needed to reconcile the discrepancies between DCNNs and human perception, particularly concerning adversarial stimuli.
Related Concept Videos
01:24Vision
59.0K
Vision is the result of light being detected and transduced into neural signals by the retina of the eye. This information is then further analyzed and interpreted by the brain. First, light enters the front of the eye and is focused by the cornea and lens onto the retina—a thin sheet of neural tissue lining the back of the eye. Because of refraction through the convex lens of the eye, images are projected onto the retina upside-down and reversed.
59.0K
01:15Depth Perception and Spatial Vision
1.5K
Depth perception is the ability to perceive objects three-dimensionally. It relies on two types of cues: binocular and monocular. Binocular cues depend on the combination of images from both eyes and how the eyes work together. Since the eyes are in slightly different positions, each eye captures a slightly different image. This disparity between images, known as binocular disparity, helps the brain interpret depth. When the brain compares these images, it determines the distance to an object.
1.5K
01:26Visual System
1.5K
Light enters the eye through the cornea, a transparent, dome-shaped surface covering the surface of the eyeball that helps to direct and focus incoming light. This light is then channeled toward the pupil, an adjustable opening whose size is controlled by the iris. The iris, a pigmented muscle, regulates the amount of light entering the eye by contracting or dilating the pupil, thereby ensuring optimal light levels for clear vision.
Once through the pupil, the light passes through the lens, a...
Once through the pupil, the light passes through the lens, a...
1.5K
01:24Color Vision
1.2K
Color perception begins in the retina, the light-sensitive layer at the back of the eye. Two main theories explain how colors are seen: the trichromatic theory and the opponent-process theory. The trichromatic theory, proposed by Thomas Young in 1802 and extended by Hermann von Helmholtz in 1852, suggests that color vision is based on three types of cone receptors in the retina. These cones are sensitive to different but overlapping ranges of wavelengths corresponding to red, blue, and green.
1.2K
01:22Photoreceptors and Visual Pathways
8.3K
At the molecular level, visual signals trigger transformations in photopigment molecules, resulting in changes in the photoreceptor cell's membrane potential. The photon's energy level is denoted by its wavelength, with each specific wavelength of visible light associated with a distinct color. The spectral range of visible light, classified as electromagnetic radiation, spans from 380 to 720 nm. Electromagnetic radiation wavelengths exceeding 720 nm fall under the infrared category,...
8.3K
01:20Anatomy of the Eyeball
9.1K
The eye is a spherical, hollow structure composed of three tissue layers. The outer layer — the fibrous tunic, comprises the sclera — a white structure — and the cornea, which is transparent. The sclera encompasses some of the ocular surface, most of which is not visible. However, the 'white of the eye' is distinctively visible in humans compared to other species. The cornea, a clear covering at the front of the eye, enables light penetration. The eye's middle...
9.1K