FWFelix A. Wichmann

Felix A. Wichmann

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Felix A. Wichmann has a total of 63 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Same stimuli, different spaces: how task, observer and amount of data influence geometry and reliability of mental representations

    by Guillermo Aguilar, Inés Schönmann, Barbara F. Mühlbauer, David-Elias Künstle, Felix Wichmann

    Similarity is central to explaining categorisation and perception in humans and artificial systems. A popular formulation treats similarity as distance in a multidimensional “mental”, “psychological” or “perceptual” space, in which similar objects lie close together. Modern ordinal embedding algorithms estimate such spaces efficiently from ordinal responses, and even though they are increasingly used to study mental representations, several methodological questions about their application and interpretation remain unresolved. Here, we address these questions with data from extensive, well-controlled psychophysical experiments. We show that popular triplet tasks are potentially more reliable than hitherto believed, but only if the substantial individual differences between observers are not discarded by pooling their responses. Furthermore, pooling across observers also inflates estimates of the dimensionality of the inferred space. To mitigate this, we propose an additional effect size criterion. Finally, we show that two commonly employed triplet tasks presumed to be comparable---classic triplets with a reference and odd-one-out tasks---are psychologically different. Even when stimuli, observers and trial-by-trial presentations are kept constant, they result in different inferred representations, often of lower dimensionality for the odd-one-out task. Spaces derived from the classic task generalise to choices in the odd-one-out task, but not vice versa, suggesting that observers might rely on impoverished decision criteria in the odd-one-out task. Our results demonstrate the potential of triplet tasks and ordinal embedding methods, but also show how seemingly innocuous design choices can impair the reliability and usefulness of the inferred mental representations.

  • Open Access

    Behavioral differences between humans and machines arise early in visual processing

    by Thomas Klein, Wieland Brendel, Felix A. Wichmann

    It remains an open question to what extent current deep neural networks (DNNs) are suitable computational models of the human visual system. While DNNs have proven to be capable of predicting neural activations in primate visual cortex with great success, psychophysical experiments have shown behavioral differences between DNNs and human observers. One of these behavioral differences is which individual images DNNs and human observers find easy or difficult to recognize, as quantified by error consistency (EC). Hypothetically, the reported differences in EC could arise late in visual processing, even though the representations extracted by DNNs and human observers may have been more similar in the initial forward sweep: At the presentation and response times investigated in earlier work, observer-internal idiosyncrasies (e.g., in feedback-mediated memory) might have influenced the final behavioral responses, lowering EC between DNNs and human observers. To test this hypothesis, we systematically vary presentation times of backward-masked stimuli from 8.3 to 267 ms and measure human performance on a speeded eightfold identification task with natural images. Contrary to the hypothesis that error consistency peaks early in time, we find that it never exceeds the value of 0.4 known from previous work with longer presentation times, suggesting that the differences between DNNs and humans cannot be explained by late high-level reasoning but point to systematic processing differences between DNNs and the early human visual system.

  • Open Access

    Estimating the contribution of early and late noise in vision from psychophysical data

    by Jesús Malo, José Juan Esteve-Taboada, Guillermo Aguilar, Marianne Maertens, Felix A. Wichmann

    Human performance in psychophysical detection and discrimination tasks is limited by inner noise. It is unclear to what extent this inner noise arises from early noise (e.g., in the photoreceptors) or from late noise (at or immediately prior to the decision stage, presumably in cortex). Very likely, the behaviorally limiting inner noise is a nontrivial combination of both early and late noise. Here we propose a method to quantify the contributions of early and late noise purely from psychophysical data. Our approach generalizes classical results for linear systems by combining the theory of noise propagation through a nonlinear network with expressions to obtain a perceptual metric through a nonlinear network. We show that from threshold-only data, the relative contributions of early and late noise can only be disentangled when the experiments include substantial external noise. When full psychometric functions are available, early and late noise sources can be quantified even in the absence of external noise. Our psychophysical estimate of the magnitude of early noise—assuming a standard cascade of linear and nonlinear model stages—is substantially lower than the noise in cone photocurrents computed via an accurate model of retinal physiology, the ISETBio. This is consistent with the idea that one of the fundamental tasks of early vision is to reduce the comparatively large retinal noise.


  • Neither hype nor gloom do DNNs justice

    by Felix A. Wichmann, Simon Kornblith, Robert Geirhos

    Neither the hype exemplified in some exaggerated claims about deep neural networks (DNNs), nor the gloom expressed by Bowers et al. do DNNs as models in vision science justice: DNNs rapidly evolve, and today's limitations are often tomorrow's successes. In addition, providing explanations as well as prediction and image-computability are model desiderata; one should not be favoured at the expense of the other.


  • Open Access

    Seeing the future of progressive glasses: a new perceptual approach to spectacle lens design

    by Yannick Sauer, David-Elias Künstle, Felix Wichmann, Siegfried Wahl

    The eye’s natural aging influences our ability to focus on close objects. Without optical correction, all adults will suffer from blurry close vision starting in their 40s. In effect, different optical corrections are necessary for near and far vision. Current state-of-the-art glasses offer a gradual change of correction across the field of view for any distance—using Progressive Addition Lenses (PALs). However, an inevitable side effect of PALs is geometric distortion, which causes the swim effect, a phenomenon of unstable perception of the environment leading to discomfort for many wearers. Unfortunately, little is known about the relationship between lens distortions and their perceptual effects, that is, between the complex physical distortions on the one hand and their subjective severity on the other.We show that perceived distortion can be measured as a psychophysical scaling function using a VR experiment with accurately simulated PAL distortions. Despite the multi-dimensional space of physical distortions, the measured perception is well represented as a 1D scaling function; distortions are perceived less with negative far correction, suggesting an advantage for short-sighted people.Beyond that, our results successfully demonstrate that psychophysical scaling with ordinal embedding methods can investigate complex perceptual phenomena like lens distortions that affect geometry, stereo, and motion perception. Our approach provides a new perspective on lens design based on modeling visual processing that could be applied beyond distortions. We anticipate that future PAL designs could be improved using our method to minimize subjectively discomforting distortions rather than merely optimizing physical parameters.

  • Open Access

    Are Deep Neural Networks Adequate Behavioral Models of Human Visual Perception?

    by Felix A. Wichmann, Robert Geirhos

    Deep neural networks (DNNs) are machine learning algorithms that have revolutionized computer vision due to their remarkable successes in tasks like object classification and segmentation. The success of DNNs as computer vision algorithms has led to the suggestion that DNNs may also be good models of human visual perception. In this article, we review evidence regarding current DNNs as adequate behavioral models of human core object recognition. To this end, we argue that it is important to distinguish between statistical tools and computational models and to understand model quality as a multidimensional concept in which clarity about modeling goals is key. Reviewing a large number of psychophysical and computational explorations of core object recognition performance in humans and DNNs, we argue that DNNs are highly valuable scientific tools but that, as of today, DNNs should only be regarded as promising—but not yet adequate—computational models of human core object recognition behavior. On the way, we dispel several myths surrounding DNNs in vision science.


  • Open Access

    The developmental trajectory of object recognition robustness: Children are like small adults but unlike big deep neural networks

    by Lukas S. Huber, Robert Geirhos, Felix A. Wichmann

    In laboratory object recognition tasks based on undistorted photographs, both adult humans and deep neural networks (DNNs) perform close to ceiling. Unlike adults’, whose object recognition performance is robust against a wide range of image distortions, DNNs trained on standard ImageNet (1.3M images) perform poorly on distorted images. However, the last 2 years have seen impressive gains in DNN distortion robustness, predominantly achieved through ever-increasing large-scale datasets—orders of magnitude larger than ImageNet. Although this simple brute-force approach is very effective in achieving human-level robustness in DNNs, it raises the question of whether human robustness, too, is simply due to extensive experience with (distorted) visual input during childhood and beyond. Here we investigate this question by comparing the core object recognition performance of 146 children (aged 4–15 years) against adults and against DNNs. We find, first, that already 4- to 6-year-olds show remarkable robustness to image distortions and outperform DNNs trained on ImageNet. Second, we estimated the number of images children had been exposed to during their lifetime. Compared with various DNNs, children’s high robustness requires relatively little data. Third, when recognizing objects, children—like adults but unlike DNNs—rely heavily on shape but not on texture cues. Together our results suggest that the remarkable robustness to distortions emerges early in the developmental trajectory of human object recognition and is unlikely the result of a mere accumulation of experience with distorted visual input. Even though current DNNs match human performance regarding robustness, they seem to rely on different and more data-hungry strategies to do so.


  • Open Access

    Are Deep Neural Networks Adequate Behavioural Models of Human Visual Perception?

    by Felix A. Wichmann, Robert Geirhos

    Deep neural networks (DNNs) are machine learning algorithms that have revolutionised computer vision due to their remarkable successes in tasks like object classification and segmentation. The success of DNNs as computer vision algorithms has led to the suggestion that DNNs may also be good models of human visual perception. We here review evidence regarding current DNNs as adequate behavioural models of human core object recognition. To this end, we argue that it is important to distinguish between statistical tools and computational models, and to understand model quality as a multidimensional concept where clarity about modelling goals is key. Reviewing a large number of psychophysical and computational explorations of core object recognition performance in humans and DNNs, we argue that DNNs are highly valuable scientific tools but that as of today DNNs should only be regarded as promising -- but not yet adequate -- computational models of human core object recognition behaviour. On the way we dispel a number of myths surrounding DNNs in vision science.

  • Open Access

    Estimating the perceived dimension of psychophysical stimuli using triplet accuracy and hypothesis testing

    by David-Elias Künstle, Ulrike von Luxburg, Felix A. Wichmann

    Vision researchers are interested in mapping complex physical stimuli to perceptual dimensions. Such a mapping can be constructed using multidimensional psychophysical scaling or ordinal embedding methods. Both methods infer coordinates that agree as much as possible with the observer’s judgments so that perceived similarity corresponds with distance in the inferred space. However, a fundamental problem of all methods that construct scalings in multiple dimensions is that the inferred representation can only reflect perception if the scale has the correct dimension. Here we propose a statistical procedure to overcome this limitation. The critical elements of our procedure are i) measuring the scale’s quality by the number of correctly predicted triplets and ii) performing a statistical test to assess if adding another dimension to the scale improves triplet accuracy significantly. We validate our procedure through extensive simulations. In addition, we study the properties and limitations of our procedure using “real” data from various behavioral datasets from psychophysical experiments. We conclude that our procedure can reliably identify (a lower bound on) the number of perceptual dimensions for a given dataset.


  • Open Access

    ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness

    by Robert Geirhos, Patricia Rubisch, Claudio Michaelis, Matthias Bethge, Felix A. Wichmann, Wieland Brendel

    Convolutional Neural Networks (CNNs) are commonly thought to recognise objects by learning increasingly complex representations of object shapes. Some recent studies suggest a more important role of image textures. We here put these conflicting hypotheses to a quantitative test by evaluating CNNs and human observers on images with a texture-shape cue conflict. We show that ImageNet-trained CNNs are strongly biased towards recognising textures rather than shapes, which is in stark contrast to human behavioural evidence and reveals fundamentally different classification strategies. We then demonstrate that the same standard architecture (ResNet-50) that learns a texture-based representation on ImageNet is able to learn a shape-based representation instead when trained on "Stylized-ImageNet", a stylized version of ImageNet. This provides a much better fit for human behavioural performance in our well-controlled psychophysical lab setting (nine experiments totalling 48,560 psychophysical trials across 97 observers) and comes with a number of unexpected emergent benefits such as improved object detection performance and previously unseen robustness towards a wide range of image distortions, highlighting advantages of a shape-based representation.