Invariant Object Representations Research Articles

Multiple Adverse Weather Conditions Adaptation for Object Detection via Causal Intervention.

Most state-of-the-art object detection methods have achieved impressive perfomrace on several public benchmarks, which are trained with high definition images. However, existing detectors are often sensitive to the visual variations and out-of-distribution data due to the domain gap caused by various confounders, e.g. the adverse weathre conditions. To bridge the gap, previous methods have been mainly exploring domain alignment, which requires to collect an amount of domain-specific training samples. In this paper, we introduce a novel domain adaptation model to discover a weather condition invariant feature representation. Specifically, we first employ a memory network to develop a confounder dictionary, which stores prototypes of object features under various scenarios. To guarantee the representativeness of each prototype in the dictionary, a dynamic item extraction strategy is used to update the memory dictionary. After that, we introduce a causal intervention reasoning module to explore the invariant representation of a specific object under different weather conditions. Finally, a categorical consistency regularization is used to constrain the similarities between categories in order to automatically search for the aligned instances among distinct domains. Experiments are conducted on several public benchmarks (RTTS, Foggy-Cityscapes, RID, and BDD 100K) with state-of-the-art performance achieved under multiple weather conditions.

The combination of Hebbian and predictive plasticity learns invariant object representations in deep sensory networks

Recognition of objects from sensory stimuli is essential for survival. To that end, sensory networks in the brain must form object representations invariant to stimulus changes, such as size, orientation and context. Although Hebbian plasticity is known to shape sensory networks, it fails to create invariant object representations in computational models, raising the question of how the brain achieves such processing. In the present study, we show that combining Hebbian plasticity with a predictive form of plasticity leads to invariant representations in deep neural network models. We derive a local learning rule that generalizes to spiking neural networks and naturally accounts for several experimentally observed properties of synaptic plasticity, including metaplasticity and spike-timing-dependent plasticity. Finally, our model accurately captures neuronal selectivity changes observed in the primate inferotemporal cortex in response to altered visual experience. Thus, we provide a plausible normative theory emphasizing the importance of predictive plasticity mechanisms for successful representational learning.

Local minimization of prediction errors drives learning of invariant object representations in a generative network model of visual perception.

The ventral visual processing hierarchy of the cortex needs to fulfill at least two key functions: perceived objects must be mapped to high-level representations invariantly of the precise viewing conditions, and a generative model must be learned that allows, for instance, to fill in occluded information guided by visual experience. Here, we show how a multilayered predictive coding network can learn to recognize objects from the bottom up and to generate specific representations via a top-down pathway through a single learning rule: the local minimization of prediction errors. Trained on sequences of continuously transformed objects, neurons in the highest network area become tuned to object identity invariant of precise position, comparable to inferotemporal neurons in macaques. Drawing on this, the dynamic properties of invariant object representations reproduce experimentally observed hierarchies of timescales from low to high levels of the ventral processing stream. The predicted faster decorrelation of error-neuron activity compared to representation neurons is of relevance for the experimental search for neural correlates of prediction errors. Lastly, the generative capacity of the network is confirmed by reconstructing specific object images, robust to partial occlusion of the inputs. By learning invariance from temporal continuity within a generative model, the approach generalizes the predictive coding framework to dynamic inputs in a more biologically plausible way than self-supervised networks with non-local error-backpropagation. This was achieved simply by shifting the training paradigm to dynamic inputs, with little change in architecture and learning rule from static input-reconstructing Hebbian predictive coding networks.

Maximal Dependence Capturing as a Principle of Sensory Processing.

Sensory inputs conveying information about the environment are often noisy and incomplete, yet the brain can achieve remarkable consistency in recognizing objects. Presumably, transforming the varying input patterns into invariant object representations is pivotal for this cognitive robustness. In the classic hierarchical representation framework, early stages of sensory processing utilize independent components of environmental stimuli to ensure efficient information transmission. Representations in subsequent stages are based on increasingly complex receptive fields along a hierarchical network. This framework accurately captures the input structures; however, it is challenging to achieve invariance in representing different appearances of objects. Here we assess theoretical and experimental inconsistencies of the current framework. In its place, we propose that individual neurons encode objects by following the principle of maximal dependence capturing (MDC), which compels each neuron to capture the structural components that contain maximal information about specific objects. We implement the proposition in a computational framework incorporating dimension expansion and sparse coding, which achieves consistent representations of object identities under occlusion, corruption, or high noise conditions. The framework neither requires learning the corrupted forms nor comprises deep network layers. Moreover, it explains various receptive field properties of neurons. Thus, MDC provides a unifying principle for sensory processing.

Which deep learning model can best explain object representations of within-category exemplars?

Deep neural network (DNN) models realize human-equivalent performance in tasks such as object recognition. Recent developments in the field have enabled testing the hierarchical similarity of object representation between the human brain and DNNs. However, the representational geometry of object exemplars within a single category using DNNs is unclear. In this study, we investigate which DNN model has the greatest ability to explain invariant within-category object representations by computing the similarity between representational geometries of visual features extracted at the high-level layers of different DNN models. We also test for the invariability of within-category object representations of these models by identifying object exemplars. Our results show that transfer learning models based on ResNet50 best explained both within-category object representation and object identification. These results suggest that the invariability of object representations in deep learning depends not on deepening the neural network but on building a better transfer learning model.

Learning Invariant Object and Spatial View Representations in the Brain Using Slow Unsupervised Learning.

First, neurophysiological evidence for the learning of invariant representations in the inferior temporal visual cortex is described. This includes object and face representations with invariance for position, size, lighting, view and morphological transforms in the temporal lobe visual cortex; global object motion in the cortex in the superior temporal sulcus; and spatial view representations in the hippocampus that are invariant with respect to eye position, head direction, and place. Second, computational mechanisms that enable the brain to learn these invariant representations are proposed. For the ventral visual system, one key adaptation is the use of information available in the statistics of the environment in slow unsupervised learning to learn transform-invariant representations of objects. This contrasts with deep supervised learning in artificial neural networks, which uses training with thousands of exemplars forced into different categories by neuronal teachers. Similar slow learning principles apply to the learning of global object motion in the dorsal visual system leading to the cortex in the superior temporal sulcus. The learning rule that has been explored in VisNet is an associative rule with a short-term memory trace. The feed-forward architecture has four stages, with convergence from stage to stage. This type of slow learning is implemented in the brain in hierarchically organized competitive neuronal networks with convergence from stage to stage, with only 4-5 stages in the hierarchy. Slow learning is also shown to help the learning of coordinate transforms using gain modulation in the dorsal visual system extending into the parietal cortex and retrosplenial cortex. Representations are learned that are in allocentric spatial view coordinates of locations in the world and that are independent of eye position, head direction, and the place where the individual is located. This enables hippocampal spatial view cells to use idiothetic, self-motion, signals for navigation when the view details are obscured for short periods.

One-shot learning of view-invariant object representations in newborn chicks

Can newborn brains perform one-shot learning? To address this question, we reared newborn chicks in strictly controlled environments containing a single view of a single object, then tested their object recognition performance across 24 uniformly-spaced viewpoints. We found that chicks can build view-invariant object representations from a single view of an object: a case of one-shot learning in newborn brains. Chicks can also build the same view-invariant object representation from different views of an object, showing that newborn brains converge on common object representations from different sets of sensory inputs. Finally, by rearing chicks with larger numbers of object views, we found that chicks develop enhanced recognition for familiar views. These results illuminate the earliest stages of object recognition, revealing (1) powerful one-shot learning that builds invariant object representations from the first views of an object and (2) view-based learning that enriches object representations, producing enhanced recognition for familiar views.

Brain-inspired automated visual object discovery and detection

Despite significant recent progress, machine vision systems lag considerably behind their biological counterparts in performance, scalability, and robustness. A distinctive hallmark of the brain is its ability to automatically discover and model objects, at multiscale resolutions, from repeated exposures to unlabeled contextual data and then to be able to robustly detect the learned objects under various nonideal circumstances, such as partial occlusion and different view angles. Replication of such capabilities in a machine would require three key ingredients: (i) access to large-scale perceptual data of the kind that humans experience, (ii) flexible representations of objects, and (iii) an efficient unsupervised learning algorithm. The Internet fortunately provides unprecedented access to vast amounts of visual data. This paper leverages the availability of such data to develop a scalable framework for unsupervised learning of object prototypes-brain-inspired flexible, scale, and shift invariant representations of deformable objects (e.g., humans, motorcycles, cars, airplanes) comprised of parts, their different configurations and views, and their spatial relationships. Computationally, the object prototypes are represented as geometric associative networks using probabilistic constructs such as Markov random fields. We apply our framework to various datasets and show that our approach is computationally scalable and can construct accurate and operational part-aware object models much more efficiently than in much of the recent computer vision literature. We also present efficient algorithms for detection and localization in new scenes of objects and their partial views.

Towards a unified perspective of object shape and motion processing in human dorsal cortex.

Although object-related areas were discovered in human parietal cortex a decade ago, surprisingly little is known about the nature and purpose of these representations, and how they differ from those in the ventral processing stream. In this article, we review evidence for the unique contribution of object areas of dorsal cortex to three-dimensional (3-D) shape representation, the localization of objects in space, and in guiding reaching and grasping actions. We also highlight the role of dorsal cortex in form-motion interaction and spatiotemporal integration, possible functional relationships between 3-D shape and motion processing, and how these processes operate together in the service of supporting goal-directed actions with objects. Fundamental differences between the nature of object representations in the dorsal versus ventral processing streams are considered, with an emphasis on how and why dorsal cortex supports veridical (rather than invariant) representations of objects to guide goal-directed hand actions in dynamic visual environments.

The Development of Invariant Object Recognition Requires Visual Experience With Temporally Smooth Objects.

How do newborns learn to recognize objects? According to temporal learning models in computational neuroscience, the brain constructs object representations by extracting smoothly changing features from the environment. To date, however, it is unknown whether newborns depend on smoothly changing features to build invariant object representations. Here, we used an automated controlled-rearing method to examine whether visual experience with smoothly changing features facilitates the development of view-invariant object recognition in a newborn animal model-the domestic chick (Gallus gallus). When newborn chicks were reared with a virtual object that moved smoothly over time, the chicks created view-invariant representations that were selective for object identity and tolerant to viewpoint changes. Conversely, when newborn chicks were reared with a temporally non-smooth object, the chicks developed less selectivity for identity features and less tolerance to viewpoint changes. These results provide evidence for a "smoothness constraint" on the development of invariant object recognition and indicate that newborns leverage the temporal smoothness of natural visual environments to build abstract mental models of objects.

Invariant object recognition is a personalized selection of invariant features in humans, not simply explained by hierarchical feed-forward vision models

One key ability of human brain is invariant object recognition, which refers to rapid and accurate recognition of objects in the presence of variations such as size, rotation and position. Despite decades of research into the topic, it remains unknown how the brain constructs invariant representations of objects. Providing brain-plausible object representations and reaching human-level accuracy in recognition, hierarchical models of human vision have suggested that, human brain implements similar feed-forward operations to obtain invariant representations. However, conducting two psychophysical object recognition experiments on humans with systematically controlled variations of objects, we observed that humans relied on specific (diagnostic) object regions for accurate recognition which remained relatively consistent (invariant) across variations; but feed-forward feature-extraction models selected view-specific (non-invariant) features across variations. This suggests that models can develop different strategies, but reach human-level recognition performance. Moreover, human individuals largely disagreed on their diagnostic features and flexibly shifted their feature extraction strategy from view-invariant to view-specific when objects became more similar. This implies that, even in rapid object recognition, rather than a set of feed-forward mechanisms which extract diagnostic features from objects in a hard-wired fashion, the bottom-up visual pathways receive, through top-down connections, task-related information possibly processed in prefrontal cortex.

Characterizing object- and position-dependent response profiles to uni- and bilateral stimulus configurations in human higher visual cortex: a 7T fMRI study

Visual scenes are initially processed via segregated neural pathways dedicated to either of the two visual hemifields. Although higher-order visual areas are generally believed to utilize invariant object representations (abstracted away from features such as stimulus position), recent findings suggest they retain more spatial information than previously thought.Here, we assessed the nature of such higher-order object representations in human cortex using high-resolution fMRI at 7T, supported by corroborative 3T data. We show that multi-voxel activation patterns in both the contra- and ipsilateral hemisphere can be exploited to successfully classify the object category of unilaterally presented stimuli. Moreover, robustly identified rank order-based response profiles demonstrated a strong contralateral bias which frequently outweighed object category preferences. Finally, we contrasted different combinatorial operations to predict the responses during bilateral stimulation conditions based on responses to their constituent unilateral elements. Results favored a max operation predominantly reflecting the contralateral stimuli. The current findings extend previous work by showing that configuration-dependent modulations in higher-order visual cortex responses as observed in single unit activity have a counterpart in human neural population coding. They furthermore corroborate the emerging view that position coding is a fundamental functional characteristic of ventral visual stream processing.

A Balanced Comparison of Object Invariances in Monkey IT Neurons.

Our ability to recognize objects across variations in size, position, or rotation is based on invariant object representations in higher visual cortex. However, we know little about how these invariances are related. Are some invariances harder than others? Do some invariances arise faster than others? These comparisons can be made only upon equating image changes across transformations. Here, we targeted invariant neural representations in the monkey inferotemporal (IT) cortex using object images with balanced changes in size, position, and rotation. Across the recorded population, IT neurons generalized across size and position both stronger and faster than to rotations in the image plane as well as in depth. We obtained a similar ordering of invariances in deep neural networks but not in low-level visual representations. Thus, invariant neural representations dynamically evolve in a temporal order reflective of their underlying computational complexity.

Contour Curvature As an Invariant Code for Objects in Visual Area V4.

Size-invariant object recognition-the ability to recognize objects across transformations of scale-is a fundamental feature of biological and artificial vision. To investigate its basis in the primate cerebral cortex, we measured single neuron responses to stimuli of varying size in visual area V4, a cornerstone of the object-processing pathway, in rhesus monkeys (Macaca mulatta). Leveraging two competing models for how neuronal selectivity for the bounding contours of objects may depend on stimulus size, we show that most V4 neurons (∼70%) encode objects in a size-invariant manner, consistent with selectivity for a size-independent parameter of boundary form: for these neurons, "normalized" curvature, rather than "absolute" curvature, provided a better account of responses. Our results demonstrate the suitability of contour curvature as a basis for size-invariant object representation in the visual cortex, and posit V4 as a foundation for behaviorally relevant object codes. Size-invariant object recognition is a bedrock for many perceptual and cognitive functions. Despite growing neurophysiological evidence for invariant object representations in the primate cortex, we still lack a basic understanding of the encoding rules that govern them. Classic work in the field of visual shape theory has long postulated that a representation of objects based on information about their bounding contours is well suited to mediate such an invariant code. In this study, we provide the first empirical support for this hypothesis, and its instantiation in single neurons of visual area V4.

The development of newborn object recognition in fast and slow visual worlds.

Object recognition is central to perception and cognition. Yet relatively little is known about the environmental factors that cause invariant object recognition to emerge in the newborn brain. Is this ability a hardwired property of vision? Or does the development of invariant object recognition require experience with a particular kind of visual environment? Here, we used a high-throughput controlled-rearing method to examine whether newborn chicks (Gallus gallus) require visual experience with slowly changing objects to develop invariant object recognition abilities. When newborn chicks were raised with a slowly rotating virtual object, the chicks built invariant object representations that generalized across novel viewpoints and rotation speeds. In contrast, when newborn chicks were raised with a virtual object that rotated more quickly, the chicks built viewpoint-specific object representations that failed to generalize to novel viewpoints and rotation speeds. Moreover, there was a direct relationship between the speed of the object and the amount of invariance in the chick's object representation. Thus, visual experience with slowly changing objects plays a critical role in the development of invariant object recognition. These results indicate that invariant object recognition is not a hardwired property of vision, but is learned rapidly when newborns encounter a slowly changing visual world.

A computational exploration of complementary learning mechanisms in the primate ventral visual pathway

In order to develop transformation invariant representations of objects, the visual system must make use of constraints placed upon object transformation by the environment. For example, objects transform continuously from one point to another in both space and time. These two constraints have been exploited separately in order to develop translation and view invariance in a hierarchical multilayer model of the primate ventral visual pathway in the form of continuous transformation learning and temporal trace learning. We show for the first time that these two learning rules can work cooperatively in the model. Using these two learning rules together can support the development of invariance in cells and help maintain object selectivity when stimuli are presented over a large number of locations or when trained separately over a large number of viewing angles.

Learning invariant object representations: asymmetric transfer of learning across line drawings and 3D cues

Learning invariant object representations: asymmetric transfer of learning across line drawings and 3D cues

Convergent and invariant object representations for sight, sound, and touch.

We continuously perceive objects in the world through multiple sensory channels. In this study, we investigated the convergence of information from different sensory streams within the cerebral cortex. We presented volunteers with three common objects via three different modalities-sight, sound, and touch-and used multivariate pattern analysis of functional magnetic resonance imaging data to map the cortical regions containing information about the identity of the objects. We could reliably predict which of the three stimuli a subject had seen, heard, or touched from the pattern of neural activity in the corresponding early sensory cortices. Intramodal classification was also successful in large portions of the cerebral cortex beyond the primary areas, with multiple regions showing convergence of information from two or all three modalities. Using crossmodal classification, we also searched for brain regions that would represent objects in a similar fashion across different modalities of presentation. We trained a classifier to distinguish objects presented in one modality and then tested it on the same objects presented in a different modality. We detected audiovisual invariance in the right temporo-occipital junction, audiotactile invariance in the left postcentral gyrus and parietal operculum, and visuotactile invariance in the right postcentral and supramarginal gyri. Our maps of multisensory convergence and crossmodal generalization reveal the underlying organization of the association cortices, and may be related to the neural basis for mental concepts.

Size precedes view: developmental emergence of invariant object representations in lateral occipital complex.

Although object perception involves encoding a wide variety of object properties (e.g., size, color, viewpoint), some properties are irrelevant for identifying the object. The key to successful object recognition is having an internal representation of the object identity that is insensitive to these properties while accurately representing important diagnostic features. Behavioral evidence indicates that the formation of these kinds of invariant object representations takes many years to develop. However, little research has investigated the developmental emergence of invariant object representations in the ventral visual processing stream, particularly in the lateral occipital complex (LOC) that is implicated in object processing in adults. Here, we used an fMR adaptation paradigm to evaluate age-related changes in the neural representation of objects within LOC across variations in size and viewpoint from childhood through early adulthood. We found a dissociation between the neural encoding of object size and object viewpoint within LOC: by age of 5-10 years, area LOC demonstrates adaptation across changes in size, but not viewpoint, suggesting that LOC responses are invariant to size variations, but that adaptation across changes in view is observed in LOC much later in development. Furthermore, activation in LOC was correlated with behavioral indicators of view invariance across the entire sample, such that greater adaptation was correlated with better recognition of objects across changes in viewpoint. We did not observe similar developmental differences within early visual cortex. These results indicate that LOC acquires the capacity to compute invariance specific to different sources of information at different time points over the course of development.

A chicken model for studying the emergence of invariant object recognition.

“Invariant object recognition” refers to the ability to recognize objects across variation in their appearance on the retina. This ability is central to visual perception, yet its developmental origins are poorly understood. Traditionally, nonhuman primates, rats, and pigeons have been the most commonly used animal models for studying invariant object recognition. Although these animals have many advantages as model systems, they are not well suited for studying the emergence of invariant object recognition in the newborn brain. Here, we argue that newly hatched chicks (Gallus gallus) are an ideal model system for studying the emergence of invariant object recognition. Using an automated controlled-rearing approach, we show that chicks can build a viewpoint-invariant representation of the first object they see in their life. This invariant representation can be built from highly impoverished visual input (three images of an object separated by 15° azimuth rotations) and cannot be accounted for by low-level retina-like or V1-like neuronal representations. These results indicate that newborn neural circuits begin building invariant object representations at the onset of vision and argue for an increased focus on chicks as an animal model for studying invariant object recognition.