A Detailed Look at Scale and Translation Invariance in a Hierarchical Neural Model of Visual Object Recognition

The HMAX model has recently been proposed by Riesenhuber & Poggio as a hierarchical model of position- and size-invariant object recognition in visual cortex. It has also turned out to model successfully a number of other properties of the ventral visual stream (the visual pathway thought to be crucial for object recognition in cortex), and particularly of (view-tuned) neurons in macaque inferotemporal cortex, the brain area at the top of the ventral stream. The original modeling study only used ``paperclip'' stimuli, as in the corresponding physiology experiment, and did not explore systematically how model units' invariance properties depended on model parameters. In this study, we aimed at a deeper understanding of the inner workings of HMAX and its performance for various parameter settings and ``natural'' stimulus classes. We examined HMAX responses for different stimulus sizes and positions systematically and found a dependence of model units' responses on stimulus position for which a quantitative description is offered. Interestingly, we find that scale invariance properties of hierarchical neural models are not independent of stimulus class, as opposed to translation invariance, even though both are affine transformations within the image plane.

Translation Invariance in Object Recognition, and Its Relation to Other Visual Transformations

Human object recognition is generally considered to tolerate changes of the stimulus position in the visual field. A number of recent studies, however, have cast doubt on the completeness of translation invariance. In a new series of experiments we tried to investigate whether positional specificity of short-term memory is a general property of visual perception. We tested same/different discrimination of computer graphics models that were displayed at the same or at different locations of the visual field, and found complete translation invariance, regardless of the similarity of the animals and irrespective of direction and size of the displacement (Exp. 1 and 2). Decisions were strongly biased towards same decisions if stimuli appeared at a constant location, while after translation subjects displayed a tendency towards different decisions. Even if the spatial order of animal limbs was randomized ("scrambled animals"), no deteriorating effect of shifts in the field of view could be detected (Exp. 3). However, if the influence of single features was reduced (Exp. 4 and 5) small but significant effects of translation could be obtained. Under conditions that do not reveal an influence of translation, rotation in depth strongly interferes with recognition (Exp. 6). Changes of stimulus size did not reduce performance (Exp. 7). Tolerance to these object transformations seems to rely on different brain mechanisms, with translation and scale invariance being achieved in principle, while rotation invariance is not.

Spatial Reference Frames for Object Recognition: Tuning for Rotations in Depth

The inferior temporal cortex (IT) of monkeys is thought to play an essential role in visual object recognition. Inferotemporal neurons are known to respond to complex visual stimuli, including patterns like faces, hands, or other body parts. What is the role of such neurons in object recognition? The present study examines this question in combined psychophysical and electrophysiological experiments, in which monkeys learned to classify and recognize novel visual 3D objects. A population of neurons in IT were found to respond selectively to such objects that the monkeys had recently learned to recognize. A large majority of these cells discharged maximally for one view of the object, while their response fell off gradually as the object was rotated away from the neuron"s preferred view. Most neurons exhibited orientation-dependent responses also during view-plane rotations. Some neurons were found tuned around two views of the same object, while a very small number of cells responded in a view- invariant manner. For five different objects that were extensively used during the training of the animals, and for which behavioral performance became view-independent, multiple cells were found that were tuned around different views of the same object. No selective responses were ever encountered for views that the animal systematically failed to recognize. The results of our experiments suggest that neurons in this area can develop a complex receptive field organization as a consequence of extensive training in the discrimination and recognition of objects. Simple geometric features did not appear to account for the neurons" selective responses. These findings support the idea that a population of neurons -- each tuned to a different object aspect, and each showing a certain degree of invariance to image transformations -- may, as an assembly, encode complex 3D objects. In such a system, several neurons may be active for any given vantage point, with a single unit acting like a blurred template for a limited neighborhood of a single view.

Learning Object-Independent Modes of Variation with Feature Flow Fields

We present a unifying framework in which "object-independent" modes of variation are learned from continuous-time data such as video sequences. These modes of variation can be used as "generators" to produce a manifold of images of a new object from a single example of that object. We develop the framework in the context of a well-known example: analyzing the modes of spatial deformations of a scene under camera movement. Our method learns a close approximation to the standard affine deformations that are expected from the geometry of the situation, and does so in a completely unsupervised (i.e. ignorant of the geometry of the situation) fashion. We stress that it is learning a "parameterization", not just the parameter values, of the data. We then demonstrate how we have used the same framework to derive a novel data-driven model of joint color change in images due to common lighting variations. The model is superior to previous models of color change in describing non-linear color changes due to lighting.

Time-Frequency Representations for Speech Signals

This work addresses two related questions. The first question is what joint time-frequency energy representations are most appropriate for auditory signals, in particular, for speech signals in sonorant regions. The quadratic transforms of the signal are examined, a large class that includes, for example, the spectrograms and the Wigner distribution. Quasi-stationarity is not assumed, since this would neglect dynamic regions. A set of desired properties is proposed for the representation: (1) shift-invariance, (2) positivity, (3) superposition, (4) locality, and (5) smoothness. Several relations among these properties are proved: shift-invariance and positivity imply the transform is a superposition of spectrograms; positivity and superposition are equivalent conditions when the transform is real; positivity limits the simultaneous time and frequency resolution (locality) possible for the transform, defining an uncertainty relation for joint time-frequency energy representations; and locality and smoothness tradeoff by the 2-D generalization of the classical uncertainty relation. The transform that best meets these criteria is derived, which consists of two-dimensionally smoothed Wigner distributions with (possibly oriented) 2-D guassian kernels. These transforms are then related to time-frequency filtering, a method for estimating the time-varying 'transfer function' of the vocal tract, which is somewhat analogous to ceptstral filtering generalized to the time-varying case. Natural speech examples are provided. The second question addressed is how to obtain a rich, symbolic description of the phonetically relevant features in these time-frequency energy surfaces, the so-called schematic spectrogram. Time-frequency ridges, the 2-D analog of spectral peaks, are one feature that is proposed. If non-oriented kernels are used for the energy representation, then the ridge tops can be identified, with zero-crossings in the inner product of the gradient vector and the direction of greatest downward curvature. If oriented kernels are used, the method can be generalized to give better orientation selectivity (e.g., at intersecting ridges) at the cost of poorer time-frequency locality. Many speech examples are given showing the performance for some traditionally difficult cases: semi-vowels and glides, nasalized vowels, consonant-vowel transitions, female speech, and imperfect transmission channels.

Rotation Invariant Real-time Face Detection and Recognition System

In this report, a face recognition system that is capable of detecting and recognizing frontal and rotated faces was developed. Two face recognition methods focusing on the aspect of pose invariance are presented and evaluated - the whole face approach and the component-based approach. The main challenge of this project is to develop a system that is able to identify faces under different viewing angles in realtime. The development of such a system will enhance the capability and robustness of current face recognition technology. The whole-face approach recognizes faces by classifying a single feature vector consisting of the gray values of the whole face image. The component-based approach first locates the facial components and extracts them. These components are normalized and combined into a single feature vector for classification. The Support Vector Machine (SVM) is used as the classifier for both approaches. Extensive tests with respect to the robustness against pose changes are performed on a database that includes faces rotated up to about 40 degrees in depth. The component-based approach clearly outperforms the whole-face approach on all tests. Although this approach isproven to be more reliable, it is still too slow for real-time applications. That is the reason why a real-time face recognition system using the whole-face approach is implemented to recognize people in color video sequences.

Learning-Based Approach to Real Time Tracking and Analysis of Faces

This paper describes a trainable system capable of tracking faces and facialsfeatures like eyes and nostrils and estimating basic mouth features such as sdegrees of openness and smile in real time. In developing this system, we have addressed the twin issues of image representation and algorithms for learning. We have used the invariance properties of image representations based on Haar wavelets to robustly capture various facial features. Similarly, unlike previous approaches this system is entirely trained using examples and does not rely on a priori (hand-crafted) models of facial features based on optical flow or facial musculature. The system works in several stages that begin with face detection, followed by localization of facial features and estimation of mouth parameters. Each of these stages is formulated as a problem in supervised learning from examples. We apply the new and robust technique of support vector machines (SVM) for classification in the stage of skin segmentation, face detection and eye detection. Estimation of mouth parameters is modeled as a regression from a sparse subset of coefficients (basis functions) of an overcomplete dictionary of Haar wavelets.

Pre-Attentive Segmentation in the Primary Visual Cortex

Stimuli outside classical receptive fields have been shown to exert significant influence over the activities of neurons in primary visual cortexWe propose that contextual influences are used for pre-attentive visual segmentation, in a new framework called segmentation without classification. This means that segmentation of an image into regions occurs without classification of features within a region or comparison of features between regions. This segmentation framework is simpler than previous computational approaches, making it implementable by V1 mechanisms, though higher leve l visual mechanisms are needed to refine its output. However, it easily handles a class of segmentation problems that are tricky in conventional methods. The cortex computes global region boundaries by detecting the breakdown of homogeneity or translation invariance in the input, using local intra-cortical interactions mediated by the horizontal connections. The difference between contextual influences near and far from region boundaries makes neural activities near region boundaries higher than elsewhere, making boundaries more salient for perceptual pop-out. This proposal is implemented in a biologically based model of V1, and demonstrated using examples of texture segmentation and figure-ground segregation. The model performs segmentation in exactly the same neural circuit that solves the dual problem of the enhancement of contours, as is suggested by experimental observations. Its behavior is compared with psychophysical and physiological data on segmentation, contour enhancement, and contextual influences. We discuss the implications of segmentation without classification and the predictions of our V1 model, and relate it to other phenomena such as asymmetry in visual search.

Observations on Cortical Mechanisms for Object Recognition andsLearning

This paper sketches a hypothetical cortical architecture for visual 3D object recognition based on a recent computational model. The view-centered scheme relies on modules for learning from examples, such as Hyperbf-like networks. Such models capture a class of explanations we call Memory-Based Models (MBM) that contains sparse population coding, memory-based recognition, and codebooks of prototypes. Unlike the sigmoidal units of some artificial neural networks, the units of MBMs are consistent with the description of cortical neurons. We describe how an example of MBM may be realized in terms of cortical circuitry and biophysical mechanisms, consistent with psychophysical and physiological data.

The Individual is Nothing, the Class Everything: Psychophysics and Modeling of Recognition in Obect Classes

Most psychophysical studies of object recognition have focussed on the recognition and representation of individual objects subjects had previously explicitely been trained on. Correspondingly, modeling studies have often employed a 'grandmother'-type representation where the objects to be recognized were represented by individual units. However, objects in the natural world are commonly members of a class containing a number of visually similar objects, such as faces, for which physiology studies have provided support for a representation based on a sparse population code, which permits generalization from the learned exemplars to novel objects of that class. In this paper, we present results from psychophysical and modeling studies intended to investigate object recognition in natural ('continuous') object classes. In two experiments, subjects were trained to perform subordinate level discrimination in a continuous object class - images of computer-rendered cars - created using a 3D morphing system. By comparing the recognition performance of trained and untrained subjects we could estimate the effects of viewpoint-specific training and infer properties of the object class-specific representation learned as a result of training. We then compared the experimental findings to simulations, building on our recently presented HMAX model of object recognition in cortex, to investigate the computational properties of a population-based object class representation as outlined above. We find experimental evidence, supported by modeling results, that training builds a viewpoint- and class-specific representation that supplements a pre-existing repre-sentation with lower shape discriminability but possibly greater viewpoint invariance.

Computational Models of Object Recognition in Cortex: A Review

Understanding how biological visual systems perform object recognition is one of the ultimate goals in computational neuroscience. Among the biological models of recognition the main distinctions are between feedforward and feedback and between object-centered and view-centered. From a computational viewpoint the different recognition tasks - for instance categorization and identification - are very similar, representing different trade-offs between specificity and invariance. Thus the different tasks do not strictly require different classes of models. The focus of the review is on feedforward, view-based models that are supported by psychophysical and physiological data.

Biologically Plausible Neural Circuits for Realization of Maximum Operations

Object recognition in the visual cortex is based on a hierarchical architecture, in which specialized brain regions along the ventral pathway extract object features of increasing levels of complexity, accompanied by greater invariance in stimulus size, position, and orientation. Recent theoretical studies postulate a non-linear pooling function, such as the maximum (MAX) operation could be fundamental in achieving such invariance. In this paper, we are concerned with neurally plausible mechanisms that may be involved in realizing the MAX operation. Four canonical circuits are proposed, each based on neural mechanisms that have been previously discussed in the context of cortical processing. Through simulations and mathematical analysis, we examine the relative performance and robustness of these mechanisms. We derive experimentally verifiable predictions for each circuit and discuss their respective physiological considerations.

Visual Segmentation without Classification in a Model of the Primary Visual Cortex

Stimuli outside classical receptive fields significantly influence the neurons' activities in primary visual cortex. We propose that such contextual influences are used to segment regions by detecting the breakdown of homogeneity or translation invariance in the input, thus computing global region boundaries using local interactions. This is implemented in a biologically based model of V1, and demonstrated in examples of texture segmentation and figure-ground segregation. By contrast with traditional approaches, segmentation occurs without classification or comparison of features within or between regions and is performed by exactly the same neural circuit responsible for the dual problem of the grouping and enhancement of contours.

Modeling Invariances in Inferotemporal Cell Tuning

In macaque inferotemporal cortex (IT), neurons have been found to respond selectively to complex shapes while showing broad tuning ("invariance") with respect to stimulus transformations such as translation and scale changes and a limited tuning to rotation in depth. Training monkeys with novel, paperclip-like objects, Logothetis et al. could investigate whether these invariance properties are due to experience with exhaustively many transformed instances of an object or if there are mechanisms that allow the cells to show response invariance also to previously unseen instances of that object. They found object-selective cells in anterior IT which exhibited limited invariance to various transformations after training with single object views. While previous models accounted for the tuning of the cells for rotations in depth and for their selectivity to a specific object relative to a population of distractor objects, the model described here attempts to explain in a biologically plausible way the additional properties of translation and size invariance. Using the same stimuli as in the experiment, we find that model IT neurons exhibit invariance properties which closely parallel those of real neurons. Simulations show that the model is capable of unsupervised learning of view-tuned neurons. The model also allows to make experimentally testable predictions regarding novel stimulus transformations and combinations of stimuli.

Generalization over contrast and mirror reversal, but not figure-ground reversal, in an "edge-based

Baylis & Driver (Nature Neuroscience, 2001) have recently presented data on the response of neurons in macaque inferotemporal cortex (IT) to various stimulus transformations. They report that neurons can generalize over contrast and mirror reversal, but not over figure-ground reversal. This finding is taken to demonstrate that ``the selectivity of IT neurons is not determined simply by the distinctive contours in a display, contrary to simple edge-based models of shape recognition'', citing our recently presented model of object recognition in cortex (Riesenhuber & Poggio, Nature Neuroscience, 1999). In this memo, I show that the main effects of the experiment can be obtained by performing the appropriate simulations in our simple feedforward model. This suggests for IT cell tuning that the possible contributions of explicit edge assignment processes postulated in (Baylis & Driver, 2001) might be smaller than expected.