Atoms of recognition in human and computer vision

By Shimon Ullman, Liav Assif, Ethan Fetaya, Daniel Harari. 2016

169 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Global-and-local attention networks for visual recognition (2018, background): “While recent advancements in DCNs have led to models that perform on par with human observers in basic visual recognition tasks, there is also growing evidence of qualitative differences in the visual strategies that they employ [47, 48, 39, 35].”
Effects of Degradations on Deep Neural Network Architectures (2018, neutral): “It was shown in [12] that even though the cropping action reduces the information content in an image, CNNs are still not comparable to HVS.”
Understanding how image quality affects deep neural networks (2016, background): “[12] consider deep neural network performance on low resolution crops of an image.”
Examining the Impact of Blur on Recognition by Convolutional Networks (2016, neutral): “[25] contrasted the drop in the accuracy of computational recognition to that of humans.”
Classification-Driven Dynamic Image Enhancement (2017, background): “[35] showed that minor changes to the image, which are barely perceptible to humans, can have drastic effects on computational recognition accuracy.”
Invariant object recognition is a personalized selection of invariant features in humans, not simply explained by hierarchical feed-forward vision models (2017, background): “It has been proposed that, the redundancy of objects’ diagnostic features can explain how invariance is achieved in recognition 28 .”
If deep learning is the answer, what is the question? (2020, background): “The recent literature is studded with comparisons of the behaviour and activity of biological and artificial systems 7–21 , summarized in a growing number of review articles 22–30 .”
Deep Nets: What have They Ever Done for Vision? (2018, background): “Recent experiments (Ullman et al., 2016) suggest that humans can interpret images unambiguously provided they are above a critical size (which depends on the image content) and additional context is unnecessary.”
What are the Visual Features Underlying Human Versus Machine Vision? (2017, methodology): “We are also thankful to Matthias Kümmerer for running DeepGaze on our images and for Danny Harari for providing MIRC stimuli used in [30].”
The Notorious Difficulty of Comparing Human and Machine Perception (2020, methodology): “One consisted of ten natural, color images whose grayscale versions were also used in the original study by Ullman et al. (2016).”