Humans, but Not Deep Neural Networks, Often Miss Giant Targets in Scenes

KOEHLER, Kathryn
Akbaş, Emre
Even with great advances in machine vision, animals are still unmatched in their ability to visually search complex scenes. Animals from bees [1, 2] to birds [3] to humans [4-12] learn about the statistical relations in visual environments to guide and aid their search for targets. Here, we investigate a novel manner in which humans utilize rapidly acquired information about scenes by guiding search toward likely target sizes. We show that humans often miss targets when their size is inconsistent with the rest of the scene, even when the targets were made larger and more salient and observers fixated the target. In contrast, we show that state-of-the-art deep neural networks do not exhibit such deficits in finding mis-scaled targets but, unlike humans, can be fooled by target-shaped distractors that are inconsistent with the expected target's size within the scene. Thus, it is not a human deficiency to miss targets when they are inconsistent in size with the scene; instead, it is a byproduct of a useful strategy that the brain has implemented to rapidly discount potential distractors.


Variability in locomotor dynamics reveals the critical role of feedback in task control.
Uyanik, I; Sefati, S; Stamper, Sa; Cho, Ka; Ankaralı, Mustafa Mert; Fortune, Es; Cowan, Nj (eLife Sciences Publications, Ltd, 2020-01-23)
Animals vary considerably in size, shape, and physiological features across individuals, but yet achieve remarkably similar behavioral performances. We examined how animals compensate for morphophysiological variation by measuring the system dynamics of individual knifefish (Eigenmannia virescens) in a refuge tracking task. Kinematic measurements of Eigenmannia were used to generate individualized estimates of each fish's locomotor plant and controller, revealing substantial variability between fish. To tes...
Widespread splicing changes in human brain development and aging
Mazin, Pavel; Xiong, Jieyi; Liu, Xiling; Yan, Zheng; Zhang, Xiaoyu; Li, Mingshuang; He, Liu; Somel, Mehmet; Yuan, Yuan; Chen, Yi-Ping Phoebe; Li, Na; Hu, Yuhui; Fu, Ning; Ning, Zhibin; Zeng, Rong; Yang, Hongyi; Chen, Wei; Gelfand, Mikhail; Khaitovich, Philipp (EMBO, 2013-01-01)
While splicing differences between tissues, sexes and species are well documented, little is known about the extent and the nature of splicing changes that take place during human or mammalian development and aging. Here, using high-throughput transcriptome sequencing, we have characterized splicing changes that take place during whole human lifespan in two brain regions: prefrontal cortex and cerebellum. Identified changes were confirmed using independent human and rhesus macaque RNA-seq data sets, exon ar...
Enhancing learning through multiple intelligences
Özdemir, Pınar; Güneysu, Sibel Çiğdem; Öztekin, Ceren (Informa UK Limited, 2006-03-01)
This study investigated whether there was a significant difference between multiple intelligence instruction (MII) and traditionally designed science instruction (TDSI) on fourth grade students' understanding of concepts associated with the 'Diversity of Living Things' unit. Students' intelligence types were also examined. There were two randomly-selected classes of 35 students of between 9 and 10 years old. The experimental group was instructed through Multiple Intelligence strategies while the control gro...
The development of narrative skills in Turkish-speaking children: A complexity approach
Ögel Balaban, Hale; Hohenberger, Annette (Public Library of Science (PLoS), 2020-5-6)
The present study examines the development of plot, evaluative and syntactic complexity in children's narratives and its relationship with gender, ToM, executive function and linguistic recursive ability. One hundred and five Turkish-speaking children distributed across 4 age groups (four-, five-, seven-eight-, and ten-eleven-year-olds) and 15 adults participated in (a) Elicitation of Narratives Task, (b) Emotional Stroop Task, (c) First- or Second-Order ToM Task (depending on their age), (d) Real-Apparent ...
Object detection through search with a foveated visual system
Akbaş, Emre (2017-10-01)
Humans and many other species sense visual information with varying spatial resolution across the visual field (foveated vision) and deploy eye movements to actively sample regions of interests in scenes. The advantage of such varying resolution architecture is a reduced computational, hence metabolic cost. But what are the performance costs of such processing strategy relative to a scheme that processes the visual field at high spatial resolution? Here we first focus on visual search and combine object det...
Citation Formats
M. P. ECKSTEIN, K. KOEHLER, L. E. WELBOURNE, and E. Akbaş, “Humans, but Not Deep Neural Networks, Often Miss Giant Targets in Scenes,” CURRENT BIOLOGY, pp. 2827–2835, 2017, Accessed: 00, 2020. [Online]. Available: