Object Recognition Challenge Research Articles

Long-distance deformation object recognition by integrating contour structure and scale-invariant heat kernel signature

The tough challenges of object recognition in long-distance scene involves contour shape deformation invariant features construction. In this work, an effective contour shape descriptor integrating critical points structure and Scale-invariant Heat Kernel Signature (SI-HKS) is proposed for long-distance object recognition. We firstly propose a general feature fusion model. Then, we capture the object contour structure feature with Critical-points Inner-distance Shape Context (CP-IDSC). Meanwhile, we pull-in the SI-HKS for capturing the local deformation-invariant properties of 2D shape. Based on the integration of the above two feature descriptors, the fusion descriptor is compacted by mapping into a low dimensional subspace using the bags-of-features, allowing for an efficient Bayesian classifier recognition. The extensive experiments on synthetic turbulence-degraded shapes and real-life infrared image show that the proposed method outperformed other compared approaches in terms of the recognition precision and robustness.

Training-less color object recognition for autonomous robotics

This paper looks at the challenge of object recognition from the perspective of achieving the final goals of practical real-world autonomous robot-vision applications; that of identifying a target of interest in the robot’s field of view and properly localizing its position in preparation for the higher-level goal of tracking and navigation. A unified framework is introduced that combines a multidimensional feature histogram approach with a multiscale pyramid approach for training-less color object recognition and localization, with direct application to autonomous robotic agents. This framework addresses the high computational cost associated with multidimensional processing by deriving the Multidimensional “Laplacian Feature Histogram Pyramid”, a novel approach to a unified multidimensional-multiscale histogram representation. Furthermore, a Taylor series formulation is employed to combine the multiscale levels of the multidimensional Laplacian feature histogram pyramid into one efficient multidimensional-multiscale “Laplacian-Taylor Feature Histogram” for rapid object recognition and localization. The paper describes the criteria for target detection and localization by autonomous robots and how this newly developed framework fits these needs. Comparative results demonstrate the robustness of this recognition framework to noise and localization of target objects in cluttered scenes.

There Is a "U" in Clutter: Evidence for Robust Sparse Codes Underlying Clutter Tolerance in Human Vision.

The neural mechanisms underlying object recognition in cluttered scenes (i.e., containing more than one object) remain poorly understood. Studies have suggested that neural responses to multiple objects correspond to an average of the responses to the constituent objects. Yet, it is unclear how the identities of an unknown number of objects could be disentangled from a confounded average response. Here, we use a popular computational biological vision model to show that averaging-like responses can result from responses of clutter-tolerant neurons to suboptimal stimuli. The model also provides a novel prediction, that human detection ability should show a U-shaped dependency on target-clutter similarity, which is confirmed experimentally, supporting a simple, unifying account of how the brain performs object recognition in clutter.

Regionlets for Generic Object Detection.

Generic object detection is confronted by dealing with different degrees of variations, caused by viewpoints or deformations in distinct object classes, with tractable computations. This demands for descriptive and flexible object representations which can be efficiently evaluated in many locations. We propose to model an object class with a cascaded boosting classifier which integrates various types of features from competing local regions, each of which may consist of a group of subregions, named as regionlets. A regionlet is a base feature extraction region defined proportionally to a detection window at an arbitrary resolution (i.e., size and aspect ratio). These regionlets are organized in small groups with stable relative positions to be descriptive to delineate fine-grained spatial layouts inside objects. Their features are aggregated into a one-dimensional feature within one group so as to be flexible to tolerate deformations. The most discriminative regionlets for each object class are selected through a boosting learning procedure. Our regionlet approach achieves very competitive performance on popular multi-class detection benchmark datasets with a single method, without any context. It achieves a detection mean average precision of 41.7 percent on the PASCAL VOC 2007 dataset, and 39.7 percent on the VOC 2010 for 20 object categories. We further develop support pixel integral images to efficiently augment regionlet features with the responses learned by deep convolutional neural networks. Our regionlet based method won second place in the ImageNet Large Scale Visual Object Recognition Challenge (ILSVRC 2013).

STDP in lateral connections creates category-based perceptual cycles for invariance learning with multiple stimuli.

Learning to recognise objects and faces is an important and challenging problem tackled by the primate ventral visual system. One major difficulty lies in recognising an object despite profound differences in the retinal images it projects, due to changes in view, scale, position and other identity-preserving transformations. Several models of the ventral visual system have been successful in coping with these issues, but have typically been privileged by exposure to only one object at a time. In natural scenes, however, the challenges of object recognition are typically further compounded by the presence of several objects which should be perceived as distinct entities. In the present work, we explore one possible mechanism by which the visual system may overcome these two difficulties simultaneously, through segmenting unseen (artificial) stimuli using information about their category encoded in plastic lateral connections. We demonstrate that these experience-guided lateral interactions robustly organise input representations into perceptual cycles, allowing feed-forward connections trained with spike-timing-dependent plasticity to form independent, translation-invariant output representations. We present these simulations as a functional explanation for the role of plasticity in the lateral connectivity of visual cortex.

Seeing different objects in different ways: Measuring ventral visual tuning to sensory and semantic features with dynamically adaptive imaging

A key challenge of object recognition is achieving a balance between selectivity for relevant features and invariance to irrelevant ones. Computational and cognitive models predict that optimal selectivity for features will differ by object, and here we investigate whether this is reflected in visual representations in the human ventral stream. We describe a new real-time neuroimaging method, dynamically adaptive imaging (DAI), that enabled measurement of neural selectivity along multiple feature dimensions in the neighborhood of single referent objects. The neural response evoked by a referent was compared to that evoked by 91 naturalistic objects using multi-voxel pattern analysis. Iteratively, the objects evoking the most similar responses were selected and presented again, to converge upon a subset that characterizes the referent's "neural neighborhood." This was used to derive the feature selectivity of the response. For three different referents, we found strikingly different selectivity, both in individual features and in the balance of tuning to sensory versus semantic features. Additional analyses placed a lower bound on the number of distinct activation patterns present. The results suggest that either the degree of specificity available for object representation in the ventral stream varies by class, or that different objects evoke different processing strategies.

Distance metric learning from uncertain side information for automated photo tagging

Automated photo tagging is an important technique for many intelligent multimedia information systems, for example, smart photo management system and intelligent digital media library. To attack the challenge, several machine learning techniques have been developed and applied for automated photo tagging. For example, supervised learning techniques have been applied to automated photo tagging by training statistical classifiers from a collection of manually labeled examples. Although the existing approaches work well for small testbeds with relatively small number of annotation words, due to the long-standing challenge of object recognition, they often perform poorly in large-scale problems. Another limitation of the existing approaches is that they require a set of high-quality labeled data, which is not only expensive to collect but also time consuming. In this article, we investigate a social image based annotation scheme by exploiting implicit side information that is available for a large number of social photos from the social web sites. The key challenge of our intelligent annotation scheme is how to learn an effective distance metric based on implicit side information (visual or textual) of social photos. To this end, we present a novel “Probabilistic Distance Metric Learning” (PDML) framework, which can learn optimized metrics by effectively exploiting the implicit side information vastly available on the social web. We apply the proposed technique to photo annotation tasks based on a large social image testbed with over 1 million tagged photos crawled from a social photo sharing portal. Encouraging results show that the proposed technique is effective and promising for social photo based annotation tasks.

Untangling invariant object recognition

Despite tremendous variation in the appearance of visual objects, primates can recognize a multitude of objects, each in a fraction of a second, with no apparent effort. However, the brain mechanisms that enable this fundamental ability are not understood. Drawing on ideas from neurophysiology and computation, we present a graphical perspective on the key computational challenges of object recognition, and argue that the format of neuronal population representation and a property that we term 'object tangling' are central. We use this perspective to show that the primate ventral visual processing stream achieves a particularly effective solution in which single-neuron invariance is not the goal. Finally, we speculate on the key neuronal mechanisms that could enable this solution, which, if understood, would have far-reaching implications for cognitive neuroscience.