We present a theory for the construction of out-of-distribution (OOD) detection features for neural networks. We introduce random features for OOD through a novel information-theoretic loss functional consisting of two terms, the first based on the KL divergence separates resulting in-distribution (ID) and OOD feature distributions and the second term is the Information Bottleneck, which favors compressed features that retain the OOD information. We formulate a variational procedure to optimize the loss and obtain OOD features. Based on assumptions on OOD distributions, one can recover properties of existing OOD features, i.e., shaping functions. Furthermore, we show that our theory can predict a new shaping function that out-performs existing ones on OOD benchmarks. Our theory provides a general framework for constructing a variety of new features with clear explainability.

We discover restrained numerical instabilities in current training practices of deep networks with stochastic gradient descent (SGD), and its variants. We show numerical error (on the order of the smallest floating point bit and thus the most extreme or limiting numerical perturbations induced from floating point arithmetic in training deep nets can be amplified significantly and result in significant test accuracy variance (sensitivities), comparable to the test accuracy variance due to stochasticity in SGD. We show how this is likely traced to instabilities of the optimization dynamics that are restrained, i.e., localized over iterations and regions of the weight tensor space. We do this by presenting a theoretical framework using numerical analysis of partial differential equations (PDE), and analyzing the gradient descent PDE of convolutional neural networks (CNNs). We show that it is stable only under certain conditions on the learning rate and weight decay. We show that rather than blowing up when the conditions are violated, the instability can be restrained. We show this is a consequence of the non-linear PDE associated with the gradient descent of the CNN, whose local linearization changes when over-driving the step size of the discretization, resulting in a stabilizing effect. We link restrained instabilities to the recently discovered Edge of Stability (EoS) phenomena, in which the stable step size predicted by classical theory is exceeded while continuing to optimize the loss and still converging. Because restrained instabilities occur at the EoS, our theory provides new insights and predictions about the EoS, in particular, the role of regularization and the dependence on the network complexity.

We empirically demonstrate numerical instabilities in training standard deep networks with SGD. Specifically, we show numerical error (on the order of the smallest floating point bit) induced from floating point arithmetic in training deep nets can be amplified significantly and result in significant test accuracy variance, comparable to the test accuracy variance due to stochasticity in SGD. We show how this is likely traced to instabilities of the optimization dynamics that are localized over iterations and regions of the weight tensor space. We do this by presenting a theoretical framework using numerical analysis of partial differential equations (PDE), and analyzing the gradient descent PDE of a one-layer convolutional neural network, which is sufficient to illustrate these instabilities. We show that it is stable only under certain conditions on the learning rate and weight decay. We reproduce the localized instabilities in the PDE for the one-layer network, which arise when the conditions are violated.

We consider the problem of optimization of cost functionals on the infinite dimensional manifold of diffeomorphisms. We present a new class of optimization methods, valid for any optimization problem setup on the space of diffeomorphisms by generalizing Nesterov accelerated optimization to the manifold of diffeomorphisms. While our framework is general for infinite dimensional manifolds, we specifically treat the case of diffeomorphisms, motivated by optical flow problems in computer vision. This is accomplished by building on a recent variational approach to a general class of accelerated optimization methods by Wibisono, Wilson, and Jordan [ Proc. Natl. Acad. Sci. USA, 113 (2016), pp. E7351--E7358], which applies in finite dimensions. We generalize that approach to infinite dimensional manifolds. We derive the surprisingly simple continuum evolution equations, which are partial differential equations, for accelerated gradient descent, and relate them to simple mechanical principles from fluid mechanics. Our approach has natural connections to the optimal mass transport problem. This is because one can think of our approach as an evolution of an infinite number of particles endowed with mass (represented with a mass density) that moves in an energy landscape. The mass evolves with the optimization variable and endows the particles with dynamics. This is different from the finite dimensional case where only a single particle moves and hence the dynamics does not depend on the mass. We derive the theory, compute the PDEs for accelerated optimization, and illustrate the behavior of these new accelerated optimization schemes.

We consider the problem of filling in missing spatiotemporal regions of a video. We provide a novel flow-based solution by introducing a generative model of images in relation to the scene (without missing regions) and mappings from the scene to images. We use the model to jointly infer the scene template, a 2D representation of the scene, and the mappings. This ensures consistency of the frame-to-frame flows generated to the underlying scene, reducing geometric distortions in flow based inpainting. The template is mapped to the missing regions in the video by a new (L2-L1) interpolation scheme, creating crisp inpaintings and reducing common blur and distortion artifacts. We show on two benchmark datasets that our approach out-performs state-of-the-art quantitatively and in user studies.1

We introduce two criteria to regularize the optimization involved in learning a classifier in a domain where no annotated data are available, leveraging annotated data in a different domain, a problem known as unsupervised domain adaptation. We focus on the task of semantic segmentation, where annotated synthetic data are aplenty, but annotating real data is laborious. The first criterion, inspired by visual psychophysics, is that the map between the two image domains be phase-preserving. This restricts the set of possible learned maps, while enabling enough flexibility to transfer semantic information. The second criterion aims to leverage ecological statistics, or regularities in the scene which are manifest in any image of it, regardless of the characteristics of the illuminant or the imaging sensor. It is implemented using a deep neural network that scores the likelihood of each possible segmentation given a single un-annotated image. Incorporating these two priors in a standard domain adaptation framework improves performance across the board in the most common unsupervised domain adaptation benchmarks for semantic segmentation.

Following the seminal work of Nesterov, accelerated optimization methods have been used to powerfully boost the performance of first-order, gradient based parameter estimation in scenarios where second-order optimization strategies are either inapplicable or impractical. Not only does accelerated gradient descent converge considerably faster than traditional gradient descent, but it also performs a more robust local search of the parameter space by initially overshooting and then oscillating back as it settles into a final configuration, thereby selecting only local minimizers with a basis of attraction large enough to contain the initial overshoot. This behavior has made accelerated and stochastic gradient search methods particularly popular within the machine learning community. In their recent PNAS 2016 paper, A Variational Perspective on Accelerated Methods in Optimization, Wibisono, Wilson, and Jordan demonstrate how a broad class of accelerated schemes can be cast in a variational framework formulated around the Bregman divergence, leading to continuum limit ODEs. We show how their formulation may be further extended to infinite dimensional manifolds (starting here with the geometric space of curves and surfaces) by substituting the Bregman divergence with inner products on the tangent space and explicitly introducing a distributed mass model which evolves in conjunction with the object of interest during the optimization process. The coevolving mass model, which is introduced purely for the sake of endowing the optimization with helpful dynamics, also links the resulting class of accelerated PDE based optimization schemes to fluid dynamical formulations of optimal mass transport.

We consider the problem of detecting objects, as they come into view, from videos in an online fashion. We provide the first real-time solution that is guaranteed to minimize the delay, i.e., the time between when the object comes in view and the declared detection time, subject to acceptable levels of detection accuracy. The method leverages modern CNN-based object detectors that operate on a single frame, to aggregate detection results over frames to provide reliable detection at a rate, specified by the user, in guaranteed minimal delay. To do this, we formulate the problem as a Quickest Detection problem, which provides the aforementioned guarantees. We derive our algorithms from this theory. We show in experiments, that with an overhead of just 50 fps, we can increase the number of correct detections and decrease the overall computational cost compared to running a modern single-frame detector.

We present SurfCut, an algorithm for extracting a smooth, simple surface with an unknown 3D curve boundary from a noisy 3D image and a seed point. Our method is built on the novel observation that ridge curves of the Euclidean length of minimal paths ending on a level set of the solution of the eikonal equation lie on the surface. Our method extracts these ridges and cuts them to form the surface boundary. Our surface extraction algorithm is built on the novel observation that the surface lies in a valley of the eikonal equation solution. The resulting surface is a collection of minimal paths. Using the framework of cubical complexes and Morse theory, we design algorithms to extract ridges and valleys robustly. Experiments on three 3D datasets show the robustness of our method, and that it achieves higher accuracy with lower computational cost than state-of-the-art.

We address the problem of texture segmentation by grouping dense pixel-wise descriptors. We introduce and construct learned Shape-Tailored Descriptors that aggregate image statistics only within regions of interest to avoid mixing statistics of different textures, and that are invariant to complex nuisances (e.g., illumination, perspective and deformations). This is accomplished by training a neural network to discriminate base shape-tailored descriptors of oriented gradients at various scales. These descriptors are defined through partial differential equations to obtain data at various scales in arbitrarily shaped regions. We formulate and optimize a joint optimization problem in the segmentation and descriptors to discriminate these base descriptors using the learned metric, equivalent to grouping learned descriptors. Experiments on benchmark datasets show that the descriptors learned on a small dataset of segmented images generalize well to unseen textures in other datasets, showing the generic nature of these descriptors. We also show state-of-the-art results on texture segmentation benchmarks.

We consider the optimization of cost functionals on manifolds and derive a variational approach to accelerated methods on manifolds. We demonstrate the methodology on the infinite-dimensional manifold of diffeomorphisms, motivated by registration problems in computer vision. We build on the variational approach to accelerated optimization by Wibisono, Wilson and Jordan, which applies in finite dimensions, and generalize that approach to infinite dimensional manifolds. We derive the continuum evolution equations, which are partial differential equations (PDE), and relate them to simple mechanical principles. Our approach can also be viewed as a generalization of the $L^2$ optimal mass transport problem. Our approach evolves an infinite number of particles endowed with mass, represented as a mass density. The density evolves with the optimization variable, and endows the particles with dynamics. This is different than current accelerated methods where only a single particle moves and hence the dynamics does not depend on the mass. We derive the theory, compute the PDEs for acceleration, and illustrate the behavior of this new accelerated optimization scheme.

We formulate an energy for segmentation that is designed to have preference for segmenting the coarse over fine structure of the image, without smoothing across boundaries of regions. The energy is formulated by integrating a continuum of scales from a scale space computed from the heat equation within regions. We show that the energy can be optimized without computing a continuum of scales, but instead from a single scale. This makes the method computationally efficient in comparison to energies using a discrete set of scales. We apply our method to texture and motion segmentation. Experiments on benchmark datasets show that a continuum of scales leads to better segmentation accuracy over discrete scales and other competing methods.

We present a general framework and method for detection of an object in a video based on apparent motion. The object moves relative to background motion at some unknown time in the video, and the goal is to detect and segment the object as soon it moves in an online manner. Due to unreliability of motion between frames, more than two frames are needed to reliably detect the object. Our method is designed to detect the object(s) with minimum delay, i.e., frames after the object moves, constraining the false alarms. Experiments on a new extensive dataset for moving object detection show that our method achieves less delay for all false alarm constraints than existing state-of-the-art.

We propose new dense descriptors for texture segmentation. Given a region of arbitrary shape in an image, these descriptors are formed from shape-dependent scale spaces of oriented gradients. These scale spaces are defined by Poisson-like partial differential equations. A key property of our new descriptors is that they do not aggregate image data across the boundary of the region, in contrast to existing descriptors based on aggregation of oriented gradients. As an example, we show how the descriptor can be incorporated in a Mumford-Shah energy for texture segmentation. We test our method on several challenging datasets for texture segmentation and textured object tracking. Experiments indicate that our descriptors lead to more accurate segmentation than non-shape dependent descriptors and the state-of-the-art in texture segmentation.

We present a new approach to wide baseline matching. We propose to use a hierarchical decomposition of the image domain and coarse-to-fine selection of regions to match. In contrast to interest point matching methods, which sample salient regions to reduce the cost of comparing all regions in two images, our method eliminates regions systematically to achieve efficiency. One advantage of our approach is that it is not restricted to covariant salient regions, which is too restrictive under large viewpoint and leads to few corresponding regions. Affine invariant matching of regions in the hierarchy is achieved efficiently by a coarse-to-fine search of the affine space. Experiments on two benchmark datasets shows that our method finds more correct correspondence of the image (with fewer false alarms) than other wide baseline methods on large viewpoint change.

We propose a method to detect disocclusion in video sequences of three-dimensional scenes and to partition the disoccluded regions into objects, defined by coherent deformation corresponding to surfaces in the scene. Our method infers deformation fields that are piecewise smooth by construction without the need for an explicit regularizer and the associated choice of weight. It then partitions the disoccluded region and groups its components with objects by leveraging on the complementarity of motion and appearance cues: Where appearance changes within an object, motion can usually be reliably inferred and used for grouping. Where appearance is close to constant, it can be used for grouping directly. We integrate both cues in an energy minimization framework, incorporate prior assumptions explicitly into the energy, and propose a numerical scheme.

We present a method to track the shape of an object from video. The method uses a joint shape and appearance model of the object, which is propagated to match shape and radiance in subsequent frames, determining object shape. Self-occlusions and dis-occlusions of the object from camera and object motion pose difficulties to joint shape and appearance models in tracking. They are unable to adapt to new shape and appearance information, leading to inaccurate shape detection. In this work, we model self-occlusions and dis-occlusions in a joint shape and appearance tracking framework. Self-occlusions and the warp to propagate the model are coupled, thus we formulate a joint optimization problem. We derive a coarse-to-fine optimization method, advantageous in tracking, that initially perturbs the model by coarse perturbations before transitioning to finer-scale perturbations seamlessly. This coarse-to-fine behavior is automatically induced by gradient descent on a novel infinite-dimensional Riemannian manifold that we introduce. The manifold consists of planar parameterized regions, and the metric that we introduce is a novel Sobolev metric. Experiments on video exhibiting occlusions/dis-occlusions, complex radiance and background show that occlusion/dis-occlusion modeling leads to superior shape accuracy.

We derive an easy-to-implement and efficient algorithm for solving multi-label image partitioning problems in the form of the problem addressed by Region Competition. These problems jointly determine a parameter for each of the regions in the partition. Given an estimate of the parameters, a fast approximate solution to the multi-label sub-problem is derived by a global update that uses smoothing and thresholding. The method is empirically validated to be robust to fine details of the image that plague local solutions. Further, in comparison to global methods for the multi-label problem, the method is more efficient and it is easy for a non-specialist to implement. We give sample Matlab code for the multi-label Chan-Vese problem in this paper. Experimental comparison to the state-of-the-art in multi-label solutions to Region Competition shows that our method achieves equal or better accuracy, with the main advantage being speed and ease of implementation.

We propose a method for tracking structures (e.g., ventricles and myocardium) in cardiac images (e.g., magnetic resonance) by propagating forward in time a previous estimate of the structures using a new physically motivated motion estimation scheme. Our method estimates motion by regularizing only within structures so that differing motions among different structures are not mixed. It simultaneously satisfies the physical constraints at the interface between a fluid and a medium that the normal component of the fluid's motion must match the normal component of the medium's motion and the No-Slip condition, which states that the tangential velocity approaches zero near the interface. We show that these conditions lead to partial differential equations with Robin boundary conditions at the interface, which couple the motion between structures. We show that propagating a segmentation across frames using our motion estimation scheme leads to more accurate segmentation than traditional motion estimation that does not use physical constraints. Our method is suited to interactive segmentation, prominently used in commercial applications for cardiac analysis, where segmentation propagation is used to predict a segmentation in the next frame. We show that our method leads to more accurate predictions than a popular and recent interactive method used in cardiac segmentation.

In this work, we address the multi-label Mumford-Shah problem, i.e., the problem of jointly estimating a partitioning of the domain of the image, and functions defined within regions of the partition. We create algorithms that are efficient, robust to undesirable local minima, and are easy-to-implement. Our algorithms are formulated by slightly modifying the underlying statistical model from which the multi-label Mumford-Shah functional is derived. The advantage of this statistical model is that the underlying variables: the labels and the functions are less coupled than in the original formulation, and the labels can be computed from the functions with more global updates. The resulting algorithms can be tuned to the desired level of locality of the solution: from fully global updates to more local updates. We demonstrate our algorithm on two applications: joint multi-label segmentation and denoising, and joint multi-label motion segmentation and flow estimation. We compare to the state-of-the-art in multi-label Mumford-Shah problems and show that we achieve more promising results.

We present a method to track the precise shape of a dynamic object in video. Joint dynamic shape and appearance models, in which a template of the object is propagated to match the object shape and radiance in the next frame, are advantageous over methods employing global image statistics in cases of complex object radiance and cluttered background. In cases of complex 3D object motion and relative viewpoint change, self-occlusions and dis-occlusions of the object are prominent, and current methods employing joint shape and appearance models are unable to accurately adapt to new shape and appearance information, leading to inaccurate shape detection. In this work, we model self-occlusions and dis-occlusions in a joint shape and appearance tracking framework. Experiments on video exhibiting occlusion/dis-occlusion, complex radiance and background show that occlusion/dis-occlusion modeling leads to superior shape accuracy compared to recent methods employing joint shape/appearance models or employing global statistics.

We define a novel metric on the space of closed planar curves which decomposes into three intuitive components. According to this metric, centroid translations, scale changes, and deformations are orthogonal, and the metric is also invariant with respect to reparameterizations of the curve. While earlier related Sobolev metrics for curves exhibit some general similarities to the novel metric proposed in this work, they lacked this important three-way orthogonal decomposition, which has particular relevance for tracking in computer vision. Another positive property of this new metric is that the Riemannian structure that is induced on the space of curves is a smooth Riemannian manifold, which is isometric to a classical well-known manifold. As a consequence, geodesics and gradients of energies defined on the space can be computed using fast closed-form formulas, and this has obvious benefits in numerical applications. The obtained Riemannian manifold of curves is ideal for addressing complex problems in computer vision; one such example is the tracking of highly deforming objects. Previous works have assumed that the object deformation is smooth, which is realistic for the tracking problem, but most have restricted the deformation to belong to a finite-dimensional group—such as affine motions—or to finitely parameterized models. This is too restrictive for highly deforming objects such as the contour of a beating heart. We adopt the smoothness assumption implicit in previous work, but we lift the restriction to finite-dimensional motions/deformations. We define a dynamical model in this Riemannian manifold of curves and use it to perform filtering and prediction to infer and extrapolate not just the pose (a finitely parameterized quantity) of an object but its deformation (an infinite-dimensional quantity) as well. We illustrate these ideas using a simple first-order dynamical model and show that it can be effective even on image sequences where existing methods fail.

We present a region-based active contour detection algorithm for objects that exhibit relatively homogeneous photometric characteristics (e.g. smooth color or gray levels), embedded in complex background clutter. Current methods either frame this problem in Bayesian classification terms, where precious modeling resources are expended representing the complex background away from decision boundaries, or use heuristics to limit the search to local regions around the object of interest. We propose an adaptive lookout region, whose size depends on the statistics of the data, that are estimated along with the boundary during the detection process. The result is a “curious snake” that explores the outside of the decision boundary only locally to the extent necessary to achieve a good tradeoff between missed detections and narrowest “lookout” region, drawing inspiration from the literature of minimum-latency set-point change detection and robust statistics. This development makes fully automatic detection in complex backgrounds a realistic possibility for active contours, allowing us to exploit their powerful geometric modeling capabilities compared with other approaches used for segmentation of cluttered scenes. To this end, we introduce an automatic initialization method tailored to our model that overcomes one of the primary obstacles in using active contours for fully automatic object detection.

This work provides a model for tubular structures, and devises an algorithm to automatically extract tubular anatomical structures from medical imagery. Our model fits many anatomical structures in medical imagery, in particular, various fiber bundles in the brain (imaged through diffusion-weighted magnetic resonance (DW-MRI)) such as the cingulum bundle, and blood vessel trees in computed tomography angiograms (CTAs). Extraction of the cingulum bundle is of interest because of possible ties to schizophrenia, and extracting blood vessels is helpful in the diagnosis of cardiovascular diseases. The tubular model we propose has advantages over many existing approaches in literature: fewer degrees-of-freedom over a general deformable surface hence energies defined on such tubes are less sensitive to undesirable local minima, and the tube (in 3-D) can be naturally represented by a 4-D curve (a radius function and centerline), which leads to computationally less costly algorithms and has the advantage that the centerline of the tube is obtained without additional effort. Our model also generalizes to tubular trees, and the extraction algorithm that we design automatically detects and evolves branches of the tree. We demonstrate the performance of our algorithm on 20 datasets of DW-MRI data and 32 datasets of CTA, and quantify the results of our algorithm when expert segmentations are available.

We consider regions of images that exhibit smooth statistics, and pose the question of characterizing the "essence" of these regions that matters for recognition. Ideally, this would be a statistic (a function of the image) that does not depend on viewpoint and illumination, and yet is sufficient for the task. In this manuscript, we show that such statistics exist. That is, one can compute deterministic functions of the image that contain all the "information" present in the original image, except for the effects of viewpoint and illumination. We also show that such statistics are supported on a "thin" (zero-measure) subset of the image domain, and thus the "information" in an image that is relevant for recognition is sparse. Yet, from this thin set one can reconstruct an image that is equivalent to the original up to a change of viewpoint and local illumination (contrast). Finally, we formalize the notion of "information" an image contains for the purpose of viewpoint- and illumination- invariant tasks, which we call "actionable information" following ideas of J. J. Gibson.

Recently proposed Sobolev active contours introduced a new paradigm for minimizing energies defined on curves by changing the traditional cost of perturbing a curve and thereby redefining their gradients. Sobolev active contours evolve more globally and are less attracted to certain intermediate local minima than traditional active contours, and it is based on a well-structured Riemannian metric. In this paper, we analyze Sobolev active contours using scale-space analysis in order to understand their evolution across different scales. This analysis shows an extremely important and useful behavior of Sobolev contours, namely, that they move successively from coarse to increasingly finer scale motions in a continuous manner. This property illustrates that one justification for using the Sobolev technique is for applications where coarse-scale deformations are preferred over fine scale deformations. Along with other properties to be discussed, the coarse-to-fine observation reveals that Sobolev active contours are, in particular, ideally suited for tracking algorithms that use active contours. We will also justify our assertion that the Sobolev metric should be used over the traditional metric for active contours in tracking problems by experimentally showing how a variety of active contour based tracking methods can be significantly improved merely by evolving the active contour according to the Sobolev method.

We present a variational method for unfolding of the cortex based on a user-chosen point of view as an alternative to more traditional global flattening methods, which incur more distortion around the region of interest. Our approach involves two novel contributions. The first is an energy function and its corresponding gradient flow to measure the average visibility of a region of interest of a surface from a given viewpoint. The second is an additional energy function and flow designed to preserve the 3D topology of the evolving surface. This latter contribution receives significant focus in this paper as it is crucial to obtain the desired unfolding effect derived from the first energy functional and flow. Without it, the resulting topology changes render the unconstrained evolution uninteresting for the purpose of cortical visualization, exploration, and inspection.

Active contour and active polygon models have been used widely for image segmentation. In some applications, the topology of the object(s) to be detected from an image is known a priori, despite a complex unknown geometry, and it is important that the active contour or polygon maintain the desired topology. In this work, we construct a novel geometric flow that can be added to image-based evolutions of active contours and polygons in order to preserve the topology of the initial contour or polygon. We emphasize that, unlike other methods for topology preservation, the proposed geometric flow continually adjusts the geometry of the original evolution in a gradual and graceful manner so as to prevent a topology change long before the curve or polygon becomes close to topology change. The flow also serves as a global regularity term for the evolving contour, and has smoothness properties similar to curvature flow. These properties of gradually adjusting the original flow and global regularization prevent geometrical inaccuracies common with simple discrete topology preservation schemes. The proposed topology preserving geometric flow is the gradient flow arising from an energy that is based on electrostatic principles. The evolution of a single point on the contour depends on all other points of the contour, which is different from traditional curve evolutions in the computer vision literature.

Recently proposed Sobolev active contours introduced a new paradigm for minimizing energies defined on curves by changing the traditional cost of perturbing a curve and thereby redefining their gradients. Sobolev active contours evolve more globally and are less attracted to certain intermediate local minima than traditional active contours. In this paper we analyze Sobolev active contours in the Fourier domain in order to understand their evolution across different scales. This analysis shows an important and useful behavior of Sobolev contours, namely, that they move successively from coarse to increasingly finer scale motions in a continuous manner. Along with other properties, the previous observation reveals that Sobolev active contours are ideally suited for tracking problems that use active contours. Our purpose in this work is to show how a variety of active contour based tracking methods can be significantly improved merely by evolving the active contours according to the Sobolev method.

Active contour and active polygon models have been used widely for image segmentation. In some applications, the topology of the object(s) to be detected from an image is known a priori, despite an unknown complex geometry, and it is important that the active contour or polygon maintain the desired topology. In this work, we construct a novel geometric flow that can be added to image based evolutions of active contours and polygons so that the topology of the initial contour or polygon is preserved. Indeed, the proposed geometric flow ensures more than just correct topology; it ensures that the active contour or polygon is, in some sense, kept far away from a topology change. Smoothness properties similar to curvature flow are also guaranteed by the proposed geometric flow. The proposed topology preserving geometric flow is the gradient flow arising from an energy that is based on electrostatic principles. The evolution of a single point on the contour depends on all other points of the contour, which is different from traditional curve evolutions in computer vision literature