MIT: Computer vision may not be as good as thought

January 24, 2008

CAMBRIDGE, Mass. - For years, scientists have been trying to teach computers how to see like humans, and recent research has seemed to show computers making progress in recognizing visual objects. A new MIT study, however, cautions that this apparent success may be misleading because the tests being used are inadvertently stacked in favor of computers.

Computer vision is important for applications ranging from "intelligent" cars to visual prosthetics for the blind. Recent computational models show apparently impressive progress, boasting 60-percent success rates in classifying natural photographic image sets. These include the widely used Caltech101 database, intended to test computer vision algorithms against the variety of images seen in the real world.

However, James DiCarlo, a neuroscientist in the McGovern Institute for Brain Research at MIT, graduate student Nicolas Pinto and David Cox of the Rowland Harvard Institute argue that these image sets have design flaws that enable computers to succeed where they would fail with more authentically varied images. For example, photographers tend to center objects in a frame and to prefer certain views and contexts. The visual system, by contrast, encounters objects in a much broader range of conditions.

"The ease with which we recognize visual objects belies the computational difficulty of this feat," explains DiCarlo, senior author of the study in the online Jan. 25 PLoS Computational Biology. "The core challenge is image variation. Any given object can cast innumerable images onto the retina depending on its position, distance, orientation, lighting and background."

The team exposed the flaws in current tests of computer object recognition by using a simple "toy" computer model inspired by the earliest steps in the brain's visual pathway. Artificial neurons with properties resembling those in the brain's primary visual cortex analyze each point in the image and capture low-level information about the position and orientation of line boundaries. The model lacks the more sophisticated analysis that happens in later stages of visual processing to extract information about higher-level features of the visual scene such as shapes, surfaces or spaces between objects.

The researchers intended this model as a straw man, expecting it to fail as a way to establish a baseline. When they tested it on the Caltech101 images, however, the model did surprisingly well, with performance similar or better than five state-of-the-art object-recognition systems.

How could that be" "We suspected that the supposedly natural images in current computer vision tests do not really engage the central problem of variability, and that our intuitions about what makes objects hard or easy to recognize are incorrect," Pinto explains.

To test this idea, the authors designed a more carefully controlled test. Using just two categories-planes and cars-they introduced variations in position, size and orientation that better reflect the range of variation in the real world.

"With only two types of objects to distinguish, this test should have been easier for the 'toy' computer model, but it proved harder," Cox says. The team's conclusion: "Our model did well on the Caltech101 image set not because it is a good model but because the 'natural' images fail to adequately capture real-world variability."

As a result, the researchers argue for revamping the current standards and images used by the computer-vision community to compare models and measure progress. Before computers can approach the performance of the human brain, they say, scientists must better understand why the task of object recognition is so difficult and the brain's abilities are so impressive.
-end-
This study was supported by the National Eye Institute, The Pew Charitable Trust and The McKnight Foundation.

Massachusetts Institute of Technology

Related Computer Model Articles from Brightsurf:

Computer model explains altered decision making in schizophrenia
Scientists have built a computer 'brain circuit', or artificial neural network, that mirrors human decision-making processes and sheds light on how circuits might be altered in psychiatric diseases.

Computer model shows how COVID-19 could lead to runaway inflammation
New study addresses a mystery first raised in March: Why do some people with COVID-19 develop severe inflammation?

Computer model predicts how drugs affect heart rhythm
UC Davis Health researchers have developed a computer model to screen drugs for unintended cardiac side effects, especially arrhythmia risk.

Computer model described the dynamic instability of microtubules
Researchers of Sechenov University together with their colleagues from several Russian institutes studied the dynamics of microtubules that form the basis of the cytoskeleton and take part in the transfer of particles within a cell and its division.

Computer model helps make sense of human memory
Researchers at the Okinawa Institute of Science and Technology Graduate University (OIST) and the RIKEN Center for Brain Science have created an artificial network to simulate the brain, demonstrating that tinkering with inhibitory circuits leads to extended memory.

Computer model could help test new sickle cell drugs
A new computer model that captures the dynamics of the red blood cell sickling process could help in evaluating drugs for treating sickle cell disease.

Novel computer model supports cancer therapy
Researchers from the Life Sciences Research Unit (LSRU) of the University of Luxembourg have developed a computer model that simulates the metabolism of cancer cells.

Reverse-engineered computer model provides new insights into larval behavior
Scientists have developed a new approach to describe the behaviors of microscopic marine larvae, which will improve future predictions of how they disperse and distribute.

New computer-aided model may help predict sepsis
Can a computer-aided model predict life-threatening sepsis? A model developed in the UK that uses routinely collected data to identify early symptoms of sepsis, published in CMAJ, shows promise.

'NarcoLogic' computer model shows unintended consequences of cocaine interdiction
Efforts to curtail the flow of cocaine into the United States from South America have made drug trafficking operations more widespread and harder to eradicate.

Read More: Computer Model News and Computer Model Current Events
Brightsurf.com is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com.