Invariant Object Recognition Research Articles

Marmoset core visual object recognition behavior is comparable to that of macaques and humans

Among the smallest simian primates, the common marmoset offers promise as an experimentally tractable primate model for neuroscience with translational potential to humans. However, given its exceedingly small brain and body, the gap in perceptual and cognitive abilities between marmosets and humans requires study. Here, we performed a comparison of marmoset behavior to that of three other species in the domain of high-level vision. We first found that marmosets outperformed rats- a marmoset-sized rodent- on a simple recognition task, with marmosets robustly recognizing objects across views. On a more challenging invariant object recognition task used previously in humans, marmosets also achieved high performance. Notably, across hundreds of images, marmosets' image-by-image behavior was highly similar to that of humans- nearly as human-like as macaque behavior. Thus, core aspects of visual perception are conserved across monkeys and humans, and marmosets present salient behavioral advantages over other small model organisms for visual neuroscience.

Sensitivity of Sparse Codes to Image Distortions.

Sparse coding has been proposed as a theory of visual cortex and as an unsupervised algorithm for learning representations. We show empirically with the MNIST data set that sparse codes can be very sensitive to image distortions, a behavior that may hinder invariant object recognition. A locally linear analysis suggests that the sensitivity is due to the existence of linear combinations of active dictionary elements with high cancellation. A nearest-neighbor classifier is shown to perform worse on sparse codes than original images. For a linear classifier with a sufficiently large number of labeled examples, sparse codes are shown to yield higher accuracy than original images, but no higher than a representation computed by a random feedforward net. Sensitivity to distortions seems to be a basic property of sparse codes, and one should be aware of this property when applying sparse codes to invariant object recognition.

The role of ventral stream areas for viewpoint-invariant object recognition

Object constancy is one of the most crucial mechanisms of the human visual system enabling viewpoint invariant object recognition. However, the neuronal foundations of object constancy are widely unknown. Research has shown that the ventral visual stream is involved in processing of various kinds of object stimuli and that several regions along the ventral stream are possibly sensitive to the orientation of an object in space. To systematically address the question of viewpoint sensitive object perception, we conducted a study with stroke patients as well as an fMRI experiment with healthy participants applying object stimuli in several spatial orientations, for example in typical and atypical viewing conditions. In the fMRI experiment, we found stronger BOLD signals and above-chance classification accuracies for objects presented in atypical viewing conditions in fusiform face sensitive and lateral occipito-temporal object preferring areas. In the behavioral patient study, we observed that lesions of the right fusiform gyrus were associated with lower performance in object recognition for atypical views. The complementary results from both experiments emphasize the contributions of fusiform and lateral-occipital areas to visual object constancy and indicate that visual object constancy is particularly enabled through increased neuronal activity and specific activation patterns for objects in demanding viewing conditions.

TARGET RECOGNITION BASED ON CENTRAL AND LOGARITHMIC CENTRAL IMAGE CHORD TRANSFORMATIONS

Object recognition is of great importance for many civil and military applications and supposes the identification and classification of the object in real-time, regardless of spatial position, angular orientation, etc. In the article, the methods of central and logarithmic central image chord transformations are described. The system for invariant object recognition and space parameters determination was developed based on the proposed image chord transformation operations, which is of a multiprocessor functionally distributed architecture, in which various modules implement separate image processing operations, and allows to recognize the objects regardless of their position, scale, or rotation, and to determine the parameters for targeting The model of a data processing flow in the system was elaborated to permit the determination of the signal processing stages which can be realized in parallel. An analytical estimation and analysis of the system’s data processing time and throughput is carrying out. The approach of the system’s optimization was applied and allows to decrease the total data processing time by 24.5 times and increase the system’s throughput by 75 times.

Using deep neural networks to evaluate object vision tasks in rats.

In the last two decades rodents have been on the rise as a dominant model for visual neuroscience. This is particularly true for earlier levels of information processing, but a number of studies have suggested that also higher levels of processing such as invariant object recognition occur in rodents. Here we provide a quantitative and comprehensive assessment of this claim by comparing a wide range of rodent behavioral and neural data with convolutional deep neural networks. These networks have been shown to capture hallmark properties of information processing in primates through a succession of convolutional and fully connected layers. We find that performance on rodent object vision tasks can be captured using low to mid-level convolutional layers only, without any convincing evidence for the need of higher layers known to simulate complex object recognition in primates. Our approach also reveals surprising insights on assumptions made before, for example, that the best performing animals would be the ones using the most abstract representations-which we show to likely be incorrect. Our findings suggest a road ahead for further studies aiming at quantifying and establishing the richness of representations underlying information processing in animal models at large.

Tactile Sensing Using Machine Learning-Driven Electrical Impedance Tomography

Electrical Impedance Tomography (EIT) tactile sensors have limited success in equipping robots with tactile sensing capabilities due to the low spatial resolution of the resulting tactile images and the presence of image artifacts. To address these limitations, we propose a modular framework for invariant recognition of objects, within the context of an EIT artificial skin sensor. Three interconnected problems, i.e., EIT image reconstruction, segmentation and object recognition, are tackled in this work with the aid of machine learning. A novel conductivity surface decomposition approach, based on low order bivariate polynomials and RBF networks is introduced for the efficient solution of the EIT inverse problem. Next, segmentation of the reconstructed images is performed using a convolutional neural network and transfer learning. Finally, a subspace KNN ensemble classifier is trained on the set of object descriptors extracted from the segmented inhomogeneities to classify the objects. The proposed framework provides an accuracy of 97.5% on unseen data.

Deep Neural Networks Point to Mid-level Complexity of Rodent Object Vision

In the last two decades rodents have been on the rise as a dominant model for visual neuroscience. This is particularly true for earlier levels of information processing, but high-profile papers have suggested that also higher levels of processing such as invariant object recognition occur in rodents. Here we provide a quantitative and comprehensive assessment of this claim by comparing a wide range of rodent behavioral and neural data with convolutional deep neural networks. These networks have been shown to capture the richness of information processing in primates through a succession of convolutional and fully connected layers. We find that rodent object vision can be captured using low to mid-level convolutional layers only, without any convincing evidence for the need of higher layers known to simulate complex object recognition in primates. Our approach also reveals surprising insights on assumptions made before, for example, that the best performing animals would be the ones using the most complex representations - which we show to likely be incorrect. Our findings suggest a road ahead for further studies aiming at quantifying and establishing the richness of representations underlying information processing in animal models at large.

Counting the learnable functions of geometrically structured data

Cover's function counting theorem is a milestone in the theory of artificial neural networks. It provides an answer to the fundamental question of determining how many binary assignments (dichotomies) of $p$ points in $n$ dimensions can be linearly realized. Regrettably, it has proved hard to extend the same approach to more advanced problems than the classification of points. In particular, an emerging necessity is to find methods to deal with structured data, and specifically with non-pointlike patterns. A prominent case is that of invariant recognition, whereby identification of a stimulus is insensitive to irrelevant transformations on the inputs (such as rotations or changes in perspective in an image). An object is therefore represented by an extended perceptual manifold, consisting of inputs that are classified similarly. Here, we develop a function counting theory for structured data of this kind, by extending Cover's combinatorial technique, and we derive analytical expressions for the average number of dichotomies of generically correlated sets of patterns. As an application, we obtain a closed formula for the capacity of a binary classifier trained to distinguish general polytopes of any dimension. These results may help extend our theoretical understanding of generalization, feature extraction, and invariant object recognition by neural networks.

Scale and translation-invariance for novel objects in human vision

Though the range of invariance in recognition of novel objects is a basic aspect of human vision, its characterization has remained surprisingly elusive. Here we report tolerance to scale and position changes in one-shot learning by measuring recognition accuracy of Korean letters presented in a flash to non-Korean subjects who had no previous experience with Korean letters. We found that humans have significant scale-invariance after only a single exposure to a novel object. The range of translation-invariance is limited, depending on the size and position of presented objects. To understand the underlying brain computation associated with the invariance properties, we compared experimental data with computational modeling results. Our results suggest that to explain invariant recognition of objects by humans, neural network models should explicitly incorporate built-in scale-invariance, by encoding different scale channels as well as eccentricity-dependent representations captured by neurons’ receptive field sizes and sampling density that change with eccentricity. Our psychophysical experiments and related simulations strongly suggest that the human visual system uses a computational strategy that differs in some key aspects from current deep learning architectures, being more data efficient and relying more critically on eye-movements.

Intrinsic dimension estimation for locally undersampled data.

Identifying the minimal number of parameters needed to describe a dataset is a challenging problem known in the literature as intrinsic dimension estimation. All the existing intrinsic dimension estimators are not reliable whenever the dataset is locally undersampled, and this is at the core of the so called curse of dimensionality. Here we introduce a new intrinsic dimension estimator that leverages on simple properties of the tangent space of a manifold and extends the usual correlation integral estimator to alleviate the extreme undersampling problem. Based on this insight, we explore a multiscale generalization of the algorithm that is capable of (i) identifying multiple dimensionalities in a dataset, and (ii) providing accurate estimates of the intrinsic dimension of extremely curved manifolds. We test the method on manifolds generated from global transformations of high-contrast images, relevant for invariant object recognition and considered a challenge for state-of-the-art intrinsic dimension estimators.

An Introduction to Hyperdimensional Computing for Robotics

Hyperdimensional computing combines very high-dimensional vector spaces (e.g. 10,000 dimensional) with a set of carefully designed operators to perform symbolic computations with large numerical vectors. The goal is to exploit their representational power and noise robustness for a broad range of computational tasks. Although there are surprising and impressive results in the literature, the application to practical problems in the area of robotics is so far very limited. In this work, we aim at providing an easy to access introduction to the underlying mathematical concepts and describe the existing computational implementations in form of vector symbolic architectures (VSAs). This is accompanied by references to existing applications of VSAs in the literature. To bridge the gap to practical applications, we describe and experimentally demonstrate the application of VSAs to three different robotic tasks: viewpoint invariant object recognition, place recognition and learning of simple reactive behaviors. The paper closes with a discussion of current limitations and open questions.

Properties of invariant object recognition in human one-shot learning suggests a hierarchical architecture different from deep convolutional neural networks

Properties of invariant object recognition in human one-shot learning suggests a hierarchical architecture different from deep convolutional neural networks

Synthesis of a hybrid brain for a humanoid robot

This article comprehends the design of a Brain-Based Robot (BBR) using hybrid techniques that incorporate both Brain-Based Device (BBD) and computational algorithms. BBDs are biologically inspired machines which have its behavior guided by a simulated nervous system. This nervous system follows detailed neuroanatomy of different brain areas. BBDs tend to have a nervous system with a large number of neurons and synapses. Thus, a huge computational power is required to simulate the nervous system of a BBD. Nevertheless, some of the tasks carried out by the simulated nervous system can be accomplished using computational algorithms which can help reducing the required computational power greatly. In this article, a BBR is built which combines some subsystems from BBD with computer vision algorithms. Computer vision algorithms are applied using OpenCV to extract some features from images, while neuronal-areas are connected together based on a detailed neuroanatomical structure to mimic the human learning process. Nengo python package is used for simulating neuronal areas in the system and monitoring activities of neuronal units. Moreover, the successful integration of the BBD’s subsystems with computer vision leads to the perceptual categorization based on invariant object-recognition of various visual cues. To make a fair comparison with BBD, the nervous system of a BBD is built on the same computer used to build the hybrid brain for the proposed BBR. The proposed hybrid brain is then applied to a Nao humanoid robot in V-REP simulation environment to test it. The results obtained through this article prove that the proposed hybrid brain possesses the same intelligence of the BBD and requires much less computational power that it can run on an on-board computer of a robot, which makes it plausible for engineering applications.

Classification and Geometry of General Perceptual Manifolds

Perceptual manifolds arise when a neural population responds to an ensemble of sensory signals associated with different physical features (e.g., orientation, pose, scale, location, and intensity) of the same perceptual object. Object recognition and discrimination requires classifying the manifolds in a manner that is insensitive to variability within a manifold. How neuronal systems give rise to invariant object classification and recognition is a fundamental problem in brain theory as well as in machine learning. Here we study the ability of a readout network to classify objects from their perceptual manifold representations. We develop a statistical mechanical theory for the linear classification of manifolds with arbitrary geometry revealing a remarkable relation to the mathematics of conic decomposition. Novel geometrical measures of manifold radius and manifold dimension are introduced which can explain the classification capacity for manifolds of various geometries. The general theory is demonstrated on a number of representative manifolds, including L2 ellipsoids prototypical of strictly convex manifolds, L1 balls representing polytopes consisting of finite sample points, and orientation manifolds which arise from neurons tuned to respond to a continuous angle variable, such as object orientation. The effects of label sparsity on the classification capacity of manifolds are elucidated, revealing a scaling relation between label sparsity and manifold radius. Theoretical predictions are corroborated by numerical simulations using recently developed algorithms to compute maximum margin solutions for manifold dichotomies. Our theory and its extensions provide a powerful and rich framework for applying statistical mechanics of linear classification to data arising from neuronal responses to object stimuli, as well as to artificial deep networks trained for object recognition tasks.

Non-accidental properties, metric invariance, and encoding by neurons in a model of ventral stream visual object recognition, VisNet.

When objects transform into different views, some properties are maintained, such as whether the edges are convex or concave, and these non-accidental properties are likely to be important in view-invariant object recognition. The metric properties, such as the degree of curvature, may change with different views, and are less likely to be useful in object recognition. It is shown that in a model of invariant visual object recognition in the ventral visual stream, VisNet, non-accidental properties are encoded much more than metric properties by neurons. Moreover, it is shown how with the temporal trace rule training in VisNet, non-accidental properties of objects become encoded by neurons, and how metric properties are treated invariantly. We also show how VisNet can generalize between different objects if they have the same non-accidental property, because the metric properties are likely to overlap. VisNet is a 4-layer unsupervised model of visual object recognition trained by competitive learning that utilizes a temporal trace learning rule to implement the learning of invariance using views that occur close together in time. A second crucial property of this model of object recognition is, when neurons in the level corresponding to the inferior temporal visual cortex respond selectively to objects, whether neurons in the intermediate layers can respond to combinations of features that may be parts of two or more objects. In an investigation using the four sides of a square presented in every possible combination, it was shown that even though different layer 4 neurons are tuned to encode each feature or feature combination orthogonally, neurons in the intermediate layers can respond to features or feature combinations present is several objects. This property is an important part of the way in which high capacity can be achieved in the four-layer ventral visual cortical pathway. These findings concerning non-accidental properties and the use of neurons in intermediate layers of the hierarchy help to emphasise fundamental underlying principles of the computations that may be implemented in the ventral cortical visual stream used in object recognition.

Quaternion polar harmonic Fourier moments for color images

This paper proposes quaternion polar harmonic Fourier moments (QPHFM) for color image processing and analyzes the properties of QPHFM. After extending Chebyshev–Fourier moments (CHFM) to quaternion Chebyshev-Fourier moments (QCHFM), comparison experiments, including image reconstruction and color image object recognition, on the performance of QPHFM and quaternion Zernike moments (QZM), quaternion pseudo-Zernike moments (QPZM), quaternion orthogonal Fourier-Mellin moments (QOFMM), QCHFM, and quaternion radial harmonic Fourier moments (QRHFM) are carried out. Experimental results show QPHFM can achieve an ideal performance in image reconstruction and invariant object recognition in noise-free and noisy conditions. In addition, this paper discusses the importance of phase information of quaternion orthogonal moments in image reconstruction.

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.

Weighted spherical Bessel–Fourier image moments

In order to improve the performance of image reconstruction and image recognition accuracy in classic image orthogonal moments, a new set of moments based on the weighted spherical Bessel polynomial of the first kind is proposed, named weighted spherical Bessel–Fourier moments (WSBFMs), which are orthogonal in polar coordinate domain and can be thought as generalized and orthogonal complex moments. Then, the set of proposed WSBFMs is derived from the weighted spherical Bessel polynomial and image rotation-invariant is easily to achieve. Compared with Zernike, orthogonal Fourier–Mellion and Bessel polynomials of the same degree, the weighted spherical Bessel orthogonal radial polynomials have more zeros value, and these zeros value are more uniformly distributed. It makes WSBFMs more suitable for geometric invariant recognition as a generalization of orthogonal complex moments. Finally, Theoretical and experimental results show the superiority of the new orthogonal moments in terms of image reconstruction capability and invariant object recognition accuracy under noise-free, noisy and smooth distortion condition.

АНАЛІЗ ВЛАСТИВОСТЕЙ, ХАРАКТЕРИСТИК ТА РЕЗУЛЬТАТІВ ЗАСТОСУВАННЯ НОВІТНІХ ДЕТЕКТОРІВ ДЛЯ ВИЗНАЧЕННЯ ОСОБЛИВИХ ТОЧОК ЗОБРАЖЕННЯ

Вирішується задача інваріантного розпізнавання візуальних об’єктів з використанням структурнихметодів на основі описів у вигляді множини особливих точок зображення. Проведено аналіз характеристикта засобів програмного моделювання сучасних методів ORB та BRISK для визначення особливих точок.Запропоновано метод бінарного аналізу для формування центрів класів та подальшої класифікації. Проведено програмне моделювання методу у порівнянні з мережею Кохонена, отримано підтвердження результативності розробленого методу для прикладної бази зображень.

Invariant recognition drives neural representations of action sequences.

Recognizing the actions of others from visual stimuli is a crucial aspect of human perception that allows individuals to respond to social cues. Humans are able to discriminate between similar actions despite transformations, like changes in viewpoint or actor, that substantially alter the visual appearance of a scene. This ability to generalize across complex transformations is a hallmark of human visual intelligence. Advances in understanding action recognition at the neural level have not always translated into precise accounts of the computational principles underlying what representations of action sequences are constructed by human visual cortex. Here we test the hypothesis that invariant action discrimination might fill this gap. Recently, the study of artificial systems for static object perception has produced models, Convolutional Neural Networks (CNNs), that achieve human level performance in complex discriminative tasks. Within this class, architectures that better support invariant object recognition also produce image representations that better match those implied by human and primate neural data. However, whether these models produce representations of action sequences that support recognition across complex transformations and closely follow neural representations of actions remains unknown. Here we show that spatiotemporal CNNs accurately categorize video stimuli into action classes, and that deliberate model modifications that improve performance on an invariant action recognition task lead to data representations that better match human neural recordings. Our results support our hypothesis that performance on invariant discrimination dictates the neural representations of actions computed in the brain. These results broaden the scope of the invariant recognition framework for understanding visual intelligence from perception of inanimate objects and faces in static images to the study of human perception of action sequences.