Acquired prosopagnosia is a rare disorder, but it serves as a model for impairments in expert-level visual processing. This review discusses five key observations made over the past 30 years. First, there are variants, an apperceptive type linked to damage to the inferior occipitotemporal cortex and an amnestic type associated with anterior temporal lesions, both either right or bilateral. Second, these variants are clustered in syndromes with other perceptual deficits, the apperceptive type with field defects, dyschromatopsia and topographagnosia, and the amnestic type with topographagnosia and the auditory disorders of phonagnosia and acquired amusia. Third, extensive testing often shows additional problems with recognizing exemplars of other objects, especially when degrees of expertise are taken into account. Fourth, the prosopagnosic impairment does not affect all facial information. For example, the perception of expression and lip-reading likely depends on other neural substrates than those for processing facial identity. Last, face perception in prosopagnosia is not immutable but can improve with extensive training, though as yet this does not represent a cure for the condition. Continuing work with neural networks and animal models will enhance our understanding of this intriguing condition and what it tells us about how our brains process vision.

Deep neural networks (DNNs) have had extraordinary successes in classifying photographic images of objects and are often described as the best models of biological vision. This conclusion is largely based on three sets of findings: (1) DNNs are more accurate than any other model in classifying images taken from various datasets, (2) DNNs do the best job in predicting the pattern of human errors in classifying objects taken from various behavioral datasets, and (3) DNNs do the best job in predicting brain signals in response to images taken from various brain datasets (e.g., single cell responses or fMRI data). However, these behavioral and brain datasets do not test hypotheses regarding what features are contributing to good predictions and we show that the predictions may be mediated by DNNs that share little overlap with biological vision. More problematically, we show that DNNs account for almost no results from psychological research. This contradicts the common claim that DNNs are good, let alone the best, models of human object recognition. We argue that theorists interested in developing biologically plausible models of human vision need to direct their attention to explaining psychological findings. More generally, theorists need to build models that explain the results of experiments that manipulate independent variables designed to test hypotheses rather than compete on making the best predictions. We conclude by briefly summarizing various promising modeling approaches that focus on psychological data.

Most conventional simultaneous localization and mapping (SLAM) approaches assume the working environment to be static. In a highly dynamic environment, this assumption divulges the impediments of a SLAM algorithm that lack modules that distinctively attend to dynamic objects despite the inclusion of optimization techniques. This work exploits such environments and reduces the effects of dynamic objects in a SLAM algorithm by separating features belonging to dynamic objects and static background using a generated binary mask image. While the features belonging to the static region are used for performing SLAM, the features belonging to non-static segments are reused instead of being eliminated. The approach employs deep neural network or DNN-based object detection module to obtain bounding boxes and then generates a lower resolution binary mask image using depth-first search algorithm over the detected semantics, characterizing the segmentation of the foreground from the static background. In addition, the features belonging to dynamic objects are tracked into consecutive frames to obtain better masking consistency. The proposed approach is tested on both publicly available dataset as well as self-collected dataset, which includes both indoor and outdoor environments. The experimental results show that the removal of features belonging to dynamic objects for a SLAM algorithm can significantly improve the overall output in a dynamic scene.

Trained features of a convolution neural network (CNN) at different convolution layers is analyzed using two quantitative metrics in this work. We first show mathematically that the Gaussian confusion measure (GCM) can be used to identify the discriminative ability of an individual feature. Next, we generalize this idea, introduce another measure called the cluster purity measure (CPM), and use it to analyze the discriminative ability of multiple features jointly. The discriminative ability of trained CNN features is validated by experimental results. Research on CNNs utilizing GCM and CPM tools offers important insights into its operational mechanism, including the behavior of trained CNN features and good detection performance of some object classes that were considered difficult in the past. Finally, the trained feature representation is compared between different CNN structures to explain the superiority of deeper networks.

Camera-based systems in dairy cattle were intensively studied over the last years. Different from this study, single camera systems with a limited range of applications were presented, mostly using 2D cameras. This study presents current steps in the development of a camera system comprising multiple 3D cameras (six Microsoft Kinect cameras) for monitoring purposes in dairy cows. An early prototype was constructed, and alpha versions of software for recording, synchronizing, sorting and segmenting images and transforming the 3D data in a joint coordinate system have already been implemented. This study introduced the application of two-dimensional wavelet transforms as method for object recognition and surface analyses. The method was explained in detail, and four differently shaped wavelets were tested with respect to their reconstruction error concerning Kinect recorded depth maps from different camera positions. The images’ high frequency parts reconstructed from wavelet decompositions using the haar and the biorthogonal 1.5 wavelet were statistically analyzed with regard to the effects of image fore- or background and of cows’ or persons’ surface. Furthermore, binary classifiers based on the local high frequencies have been implemented to decide whether a pixel belongs to the image foreground and if it was located on a cow or a person. Classifiers distinguishing between image regions showed high (⩾0.8) values of Area Under reciever operation characteristic Curve (AUC). The classifications due to species showed maximal AUC values of 0.69.

This paper presents a method to improve the speed and accuracy rate for space robot visual target recognition based on illumination and affine invariant feature extraction. The method takes illumination changes, strong nonlinear light due to refraction and reflection, target affine transformation and occlusion into consideration, all of which occur on the cabin target surface and affect the target recognition accuracy seriously. In this paper, a method is proposed to capture a same target at multi-viewpoints to establish feature library for high recognition accuracy and speed at any viewpoint. By using an analysis of the light intensity and gray level transformation, we obtain the corrected image which reduce the influence of illumination change. Then the affine moment invariants features of the correction images at multi-viewpoints were extracted and the average feature datum were stored in the library. To verify the validity of the method, a robot vision system provided images, followed by image preprocessing, dynamic local threshold segmentation and feature extraction. These methods were verified on a target recognition system of space robot built for this research. The experimental results showed that the methods were feasible and effective.

Purely parallel neural networks can model object recognition in brief displays – the same conditions under which illusory conjunctions (the incorrect combination of features into perceived objects in a stimulus array) have been demonstrated empirically (Treisman 1986; Treisman & Gelade 1980). Correcting errors of illusory conjunction is the “tag-assignment” problem for a purely parallel processor: the problem of assigning a spatial tag to nonspatial features, feature combinations, and objects. This problem must be solved to model human object recognition over a longer time scale. Our model simulates both the parallel processes that may underlie illusory conjunctions and the serial processes that may solve the tag-assignment problem in normal perception. One component of the model extracts pooled features and another provides attentional tags that correct illusory conjunctions. Our approach addresses two questions: (i) How can objects be identified from simultaneously attended features in a parallel, distributed representation? (ii) How can the spatial selectional requirements of such an attentional process be met by a separation of pathways for spatial and nonspatial processing? Our analysis of these questions yields a neurally plausible simulation of tag assignment based on synchronizing feature processing activity in a spatial focus of attention.

Fundamental to spatial knowledge in all species are the representations underlying object recognition, object search, and navigation through space. But what sets humans apart from other species is our ability to express spatial experience through language. This target article explores the language of objects and places, asking what geometric properties are preserved in the representations underlying object nouns and spatial prepositions in English. Evidence from these two aspects of language suggests there are significant differences in the geometric richness with which objects and places are encoded. When an object is named (i.e., with count nouns), detailed geometric properties – principally the object's shape (axes, solid and hollow volumes, surfaces, and parts) – are represented. In contrast, when an object plays the role of either “figure” (located object) or “ground” (reference object) in a locational expression, only very coarse geometric object properties are represented, primarily the main axes. In addition, the spatial functions encoded by spatial prepositions tend to be nonmetric and relatively coarse, for example, “containment,” “contact,” “relative distance,” and “relative direction.” These properties are representative of other languages as well. The striking differences in the way language encodes objects versus places lead us to suggest two explanations: First, there is a tendency for languages to level out geometric detail from both object and place representations. Second, a nonlinguistic disparity between the representations of “what” and “where” underlies how language represents objects and places. The language of objects and places converges with and enriches our understanding of corresponding spatial representations.

Most studies on people with age-related macular degeneration (AMD) have been focused on investigations of low-level processes with simple stimuli like gratings, letters, and in perception of isolated faces or objects. We investigated the ability of people with low vision to analyze more complex stimuli like photographs of natural scenes. Fifteen participants with AMD and low vision (acuity on the better eye <20/200) and 11 normally sighted age-matched controls took part in the study. They were presented with photographs of either colored or achromatic gray level scenes in one condition and with photographs of natural scenes versus isolated objects extracted from these scenes in another condition. The photographs were centrally displayed for 300 ms. In both conditions, observers were instructed to press a key when they saw a predefined target (a face or an animal). The target was present in half of the trials. Color facilitated performance in people with low vision, while equivalent performance was found for colored and achromatic pictures in normally sighted participants. Isolated objects were categorized more accurately than objects in scenes in people with low vision. No difference was found for normally sighted observers. The results suggest that spatial properties that facilitate image segmentation (e.g., color and reduced crowding) help object perception in people with low vision.

Category-specific impairments of object recognition and naming are among the most intriguing disorders in neuropsychology, affecting the retrieval of knowledge about either living or nonliving things. They can give us insight into the nature of our representations of objects: Have we evolved different neural systems for recognizing different categories of object? What kinds of knowledge are important for recognizing particular objects? How does visual similarity within a category influence object recognition and representation? What is the nature of our semantic knowledge about different objects? We review the evidence on category-specific impairments, arguing that deficits even for one class of object (e.g., living things) cannot be accounted for in terms of a single information processing disorder across all patients; problems arise at contrasting loci in different patients. The same apparent pattern of impairment can be produced by damage to different loci. According to a new processing framework for object recognition and naming, the hierarchical interactive theory (HIT), we have a hierarchy of highly interactive stored representations. HIT explains the variety of patients in terms of (1) lesions at different levels of processing and (2) different forms of stored knowledge used both for particular tasks and for particular categories of object.

Event-related potentials were recorded using color pictures of real objects. Participants made relatedness judgments for pictures that were highly, moderately, or unrelated to a picture of a preceding prime object (Experiment 1) or object identification decisions for related/easily identified, unrelated/easily identified, and unrelated/unidentifiable objects preceded by prime objects (Experiment 2). Unrelated pictures elicited larger event-related potential negativities between 225 and 500 ms than did related pictures, although the first portion of this epoch had a more frontal distribution than did the later portion. The later epoch differentiated the unrelated from the moderately related and the moderately related from the highly related pictures (Experiment 1), but the early epoch produced differences only between the unrelated and related pictures (Experiments 1 and 2). This pattern supports the existence of two separate components, an anterior, image-specific N300 and a later, central/parietal amodal N400.

A common problem in Bayesian object recognition using marked point process models is to produce a point estimate of the true underlying object configuration: the number of objects and the size, location and shape of each object. We use decision theory and the concept of loss functions to design a more reasonable estimator for this purpose, rather than using the common zero-one loss corresponding to the maximum a posteriori estimator. We propose to use the squared Δ-metric of Baddeley (1992) as our loss function and demonstrate that the corresponding optimal Bayesian estimator can be well approximated by combining Markov chain Monte Carlo methods with simulated annealing into a two-step algorithm. The proposed loss function is tested using a marked point process model developed for locating cells in confocal microscopy images.

It is worthwhile to search for forms of coding, processing, and learning common to various cortical regions and cognitive functions. Local cortical processors may coordinate their activity by maximizing the transmission of information coherently related to the context in which it occurs, thus forming synchronized population codes. This coordination involves contextual field (CF) connections that link processors within and between cortical regions. The effects of CF connections are distinguished from those mediating receptive field (RF) input; it is shown how CFs can guide both learning and processing without becoming confused with the transmission of RF information. Simulations explore the capabilities of networks built from local processors with both RF and CF connections. Physiological evidence for synchronization, CFs, and plasticity of the RF and CF connections is described. Coordination via CFs is related to perceptual grouping, the effects of context on contrast sensitivity, amblyopia, implicit influences of color in achromotopsia, object and word perception, and the discovery of distal environmental variables and their interactions through self-organization. Cortical computation could thus involve the flexible evaluation of relations between input signals by locally specialized but adaptive processors whose activity is dynamically associated and coordinated within and between regions through specialized contextual connections.

There have been several proposals for the mechanism by which we are able to recognize an object across a number of viewpoints. Viewpoint-dependent accounts suggest that recognition may be based on an incremental transformation (e.g., mental rotation) strategy, while a variety of viewpoint-independent mechanisms for object recognition have also been proposed. Recent research in neurobiology, based on the two cortical visual systems account, suggest that the processes of viewpoint-dependent and viewpoint-independent object recognition may rely on separate anatomical regions, and that brain lesions may leave patients with selective access to particular types of representation. Evidence from a variety of neuropsychological disorders are reviewed to support the position that viewpoint-independent object recognition depends upon the integrity of occipitotemporal structures. In addition, it is suggested that viewpoint-dependent processes (perhaps depending on occipitoparietal structures) may supplement this primary system under nonoptimal circumstances. (JINS, 1997, 3, 288–298.)