The visual and semantic features that predict object memory: Concept property norms for 1,000 object images

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computer Vision

Background:

• Human visual long-term memory exhibits high fidelity, yet its composition remains debated.
• Research increasingly focuses on visual object recognition and stimulus memorability.
• Object representation is viewed as a hierarchy from low-level visual to high-level conceptual organization.

Purpose of the Study:

• Investigate the contributions of visual and semantic factors to different forms of everyday object memory.
• Examine how high- and low-level features interact to influence memory distinctiveness.
• Provide a comprehensive understanding of factors affecting object memorability.

Main Methods:

• Collected normative visual and semantic feature data for 1,000 object images.
• Conducted a memory study with picture encoding on Day 1.
• Administered Lexical (lexical cue) or Visual (picture cue) memory tests on Day 2.

Main Results:

• Higher-level visual factors (Deep Neural Networks) and semantic factors (feature statistics) contribute independently to object memory.
• Semantic information influences both true and false memory performance.
• Predictors of object memory vary depending on the specific memory test used.

Conclusions:

• Visual and semantic features play distinct roles in object memory formation and retrieval.
• Semantic knowledge impacts the accuracy and susceptibility to false memories.
• Understanding these factors is crucial for a complete model of visual long-term memory.