Related Experiment Video
Updated: Jul 16, 2026

06:14
A Method for Investigating Change Blindness in Pigeons (Columba Livia)
Published on: September 7, 2018
Pigeons' recognition of partially occluded objects depends on specific training experience
Olga F Lazareva1, Edward A Wasserman, Irving Biederman
1Department of Psychology, E11 SSH, University of Iowa, Iowa City, IA 52242-1407, USA. olga-lazareva@uiowa.edu
Perception
|March 16, 2007
Summary
Prior training significantly impacts pigeons' ability to recognize partially occluded objects. While notched object training helped, it did not generalize to new items, unlike previous on-top training methods.
Area of Science:
- Cognitive psychology
- Animal behavior
- Visual perception
Background:
- Pigeons demonstrate remarkable object recognition capabilities.
- Previous research showed training with objects on a surface improved recognition of occluded items.
- The role of specific training methods in enhancing object recognition remains under investigation.
Purpose of the Study:
- To determine if training with partially erased or notched stimuli enhances pigeons' recognition of partially occluded objects.
- To compare the effects of erased and notched training with prior 'on-top' training findings.
- To understand how prior experience influences the decomposition of complex visual scenes in pigeons.
Main Methods:
- Pigeons were trained to recognize objects presented with partial erasures.
- Pigeons were trained to recognize objects with a thin gap (notches) between the object and a surface.
- Performance on recognizing partially occluded versions of trained and novel objects was assessed.
Main Results:
- Training with partially erased objects did not improve recognition of partially occluded objects.
- Training with notched objects improved recognition of the same partially occluded objects.
- The improvement from notched training did not transfer to novel partially occluded objects.
Conclusions:
- Prior experience is crucial for pigeons' recognition of partially occluded objects.
- Training methods that enhance the decomposition of complex visual scenes improve occluded object recognition.
- Notched training offers some benefit but lacks the generalizability of on-top training for novel object recognition.

