Plant behavior and response to environmental stimuli has tremendous importance in science and agriculture. In particular, a plant's root continuously senses changes in the environment, and responds in ways that optimize dynamically different essential parameters like its stability, and adequate food and water supplies. Some of the plant behavioral changes in response to environmental changes, like water shortage, can be reversible, and after certain “stress” time, the plant can get back to its normal behavioral patterns. In other cases, the plant behavior after the stress stimulus ends, is changed, due to effects on internal mechanisms, facilitating long-term behavioral changes. The main aim of this work is to derive a preliminary physical model and analysis tools to quantify the behavioral changes of a plant in response to a stimuli. To demonstrate the model, we examined, without loss of generality, the change in plant growth rate in response to electrical simuli. We showed how the suggested plant behavioral model can assist in computational analysis of short and long term plant response to changing stimuli, construct a common baseline for comparison with other stimuli, and derive new quantitative measurements that can be correlated with internal plant mechanism and assist in assessing behavioral plant patterns and in the design of more efficient agricultural technologies.

A commonly used learning rule is to approximately minimize the \emphaverage loss over the training set. Other learning algorithms, such as AdaBoost and hard-SVM, aim at minimizing the \emphmaximal loss over the training set. The average loss is more popular, particularly in deep learning, due to three main reasons. First, it can be conveniently minimized using online algorithms, that process few examples at each iteration. Second, it is often argued that there is no sense to minimize the loss on the training set too much, as it will not be reflected in the generalization loss. Last, the maximal loss is not robust to outliers. In this paper we describe and analyze an algorithm that can convert any online algorithm to a minimizer of the maximal loss. We show, theoretically and empirically, that in some situations better accuracy on the training set is crucial to obtain good performance on unseen examples. Last, we propose robust versions of the approach that can handle outliers.

This work is motivated by the engineering task of achieving a near state-of-the-art face recognition on a minimal computing budget running on an embedded system. Our main technical contribution centers around a novel training method, called Multibatch, for similarity learning, i.e., for the task of generating an invariant ``face signature'' through training pairs of ``same'' and ``not-same'' face images. The Multibatch method first generates signatures for a mini-batch of $k$ face images and then constructs an unbiased estimate of the full gradient by relying on all $k^2-k$ pairs from the mini-batch. We prove that the variance of the Multibatch estimator is bounded by $O(1/k^2)$, under some mild conditions. In contrast, the standard gradient estimator that relies on random $k/2$ pairs has a variance of order $1/k$. The smaller variance of the Multibatch estimator significantly speeds up the convergence rate of stochastic gradient descent. Using the Multibatch method we train a deep convolutional neural network that achieves an accuracy of $98.2\%$ on the LFW benchmark, while its prediction runtime takes only $30$msec on a single ARM Cortex A9 core. Furthermore, the entire training process took only 12 hours on a single Titan X GPU.

Multiclass prediction is the problem of classifying an object into a relevant target class. We consider the problem of learning a multiclass predictor that uses only few features, and in particular, the number of used features should increase sub-linearly with the number of possible classes. This implies that features should be shared by several classes. We describe and analyze the ShareBoost algorithm for learning a multiclass predictor that uses few shared features. We prove that ShareBoost efficiently finds a predictor that uses few shared features (if such a predictor exists) and that it has a small generalization error. We also describe how to use ShareBoost for learning a non-linear predictor that has a fast evaluation time. In a series of experiments with natural data sets we demonstrate the benefits of ShareBoost and evaluate its success relatively to other state-of-the-art approaches.

We present a novel image operator that seeks to find the value of stroke width for each image pixel, and demonstrate its use on the task of text detection in natural images. The suggested operator is local and data dependent, which makes it fast and robust enough to eliminate the need for multi-scale computation or scanning windows. Extensive testing shows that the suggested scheme outperforms the latest published algorithms. Its simplicity allows the algorithm to detect texts in many fonts and languages.

We present a method for efficient and reliable geo-positioning of images. It relies on image-based matching of the query images onto a trellis of existing images that provides accurate 5-DOF calibration (camera position and orientation without scale). As such it can handle any image input, including old historical images, matched against a whole city. On such a scale, care needs to be taken with the size of the database. We deviate from previous work by using 360° panoramas to simultaneously reduce the database size and increase the coverage. To reduce the likelihood of false matches, we restrict the range of angles for matched features. Furthermore, we enhance the RANSAC procedure to include two phases. The second phase includes guided feature matching to increase the likelihood of positive matches. Hence, we devise a matching confidence score that separates between true and false matches. We demonstrate the algorithm on a large scale database covering a whole city in order to show its usefulness for a vision-based augmented reality system.

Abstract We present an automatic method to recover high‐resolution texture over an object by mapping detailed photographs onto its surface. Such high‐resolution detail often reveals inaccuracies in geometry and registration, as well as lighting variations and surface reflections. Simple image projection results in visible seams on the surface. We minimize such seams using a global optimization that assigns compatible texture to adjacent triangles. The key idea is to search not only combinatorially over the source images, but also over a set of local image transformations that compensate for geometric misalignment. This broad search space is traversed using a discrete labeling algorithm, aided by a coarse‐to‐fine strategy. Our approach significantly improves resilience to acquisition errors, thereby allowing simple and easy creation of textured models for use in computer graphics.

We reduce transmission bandwidth and memory space for images by factoring their repeated content. A transform map and a condensed epitome are created such that all image blocks can be reconstructed from transformed epitome patches. The transforms may include affine deformation and color scaling to account for perspective and tonal variations across the image. The factored representation allows efficient random-access through a simple indirection, and can therefore be used for real-time texture mapping without expansion in memory. Our scheme is orthogonal to traditional image compression, in the sense that the epitome is amenable to further compression such as DXT. Moreover it allows a new mode of progressivity, whereby generic features appear before unique detail. Factoring is also effective across a collection of images, particularly in the context of image-based rendering. Eliminating redundant content lets us include textures that are several times as large in the same memory space.

We present a novel framework for organizing large collections of images in a hierarchical way, based on scene semantics. Rather than score images directly, we use them to score the scene in order to identify typical views and important locations which we term Geo-Relevance. This is done by relating each image with its viewing frustum which can be readily computed for huge collections of images nowadays. The frustum contains much more information than only camera position that has been used so far. For example, it distinguishes between a photo of the Eiffel Tower and a photo of a garbage bin taken from the exact same place. The proposed framework enables a summarized display of the information and facilitates efficient browsing.

This paper presents a new framework for the completion of missing information based on local structures. It poses the task of completion as a global optimization problem with a well-defined objective function and derives a new algorithm to optimize it. Missing values are constrained to form coherent structures with respect to reference examples. We apply this method to space-time completion of large space-time "holes" in video sequences of complex dynamic scenes. The missing portions are filled in by sampling spatio-temporal patches from the available parts of the video, while enforcing global spatio-temporal consistency between all patches in and around the hole. The consistent completion of static scene parts simultaneously with dynamic behaviors leads to realistic looking video sequences and images. Space-time video completion is useful for a variety of tasks, including, but not limited to: 1) Sophisticated video removal (of undesired static or dynamic objects) by completing the appropriate static or dynamic background information. 2) Correction of missing/corrupted video frames in old movies. 3) Modifying a visual story by replacing unwanted elements. 4) Creation of video textures by extending smaller ones. 5) Creation of complete field-of-view stabilized video. 6) As images are one-frame videos, we apply the method to this special case as well.

The space of images is known to be a nonlinear sub-space that is difficult to model. This paper derives an algorithm that walks within this space. We seek a manifold through the video volume that is constrained to lie locally in this space. Every local neighborhood within the manifold resembles some image patch. We call this the scene manifold because the solution traces the scene outline. For a broad class of inputs the problem can be posed as finding the shortest path in a graph and can thus be solved efficiently to produce the globally optimal solution. Constraining appearance rather than geometry gives rise to numerous new capabilities. Here we demonstrate the usefulness of this approach by posing the well-studied problem of mosaicing in a new way. Instead of treating it as geometrical alignment, we pose it as an appearance optimization. Since the manifold is constrained to lie in the space of valid image patches, the resulting mosaic is guaranteed to have the least distortions possible. Any small part of it can be seen in some image even though the manifold spans the whole video. Thus it can deal seamlessly with both static and dynamic scenes, with or without 3D parallax. Essentially, the method simultaneously solves two problems that have been solved only separately until now: alignment and mosaicing.

We wish to determine the epipolar geometry of a stereo camera pair from image measurements alone. This paper describes a solution to this problem, which does not require a parametric model of the camera system, and consequently applies equally well to a wide class of stereo configurations. Examples in the paper range from a standard pinhole stereo configuration to more exotic systems combining curved mirrors and wide-angle lenses. The method described here allows epipolar curves to be learnt from multiple image pairs acquired by stereo cameras with fixed configuration. By aggregating information over the multiple image pairs, a dense map of the epipolar curves can be determined on the images. The algorithm requires a large number of images, but has the distinct benefit that the correspondence problem does not have to be explicitly solved. We show that for standard stereo configurations the results are comparable to those obtained from a state of the art parametric model method, despite the significantly weaker constraints on the non-parametric model. The new algorithm is simple to implement, so it may easily be employed on a new and possibly complex camera system.

Given a set of images acquired from known viewpoints, we describe a method for synthesizing the image which would be seen from a new viewpoint. In contrast to existing techniques, which explicitly reconstruct the 3D geometry of the scene, we transform the problem to the reconstruction of colour rather than depth. This retains the benefits of geometric constraints, but projects out the ambiguities in depth estimation which occur in textureless regions. On the other hand, regularization is still needed in order to generate high-quality images. The paper's second contribution is to constrain the generated views to lie in the space of images whose texture statistics are those of the input images. This amounts to an image-based prior on the reconstruction which regularizes the solution, yielding realistic synthetic views. Examples are given of new view generation for cameras interpolated between the acquisition viewpoints - which enables synthetic steadicam stabilization of a sequence with a high level of realism.

This paper discusses two efficient methods for image based rendering. Both algorithms approximate the world object. One uses the convex hull and the other uses the visual hull. We show that the overhead of using the latter is not always justified and in some cases, might even hurt. We demonstrate the method on real images from a studio-like setting in which many cameras are used, after a simple calibration procedure. The novelties of this paper include showing that projective calibration suffices for this computation and providing simpler formulation that is base only on image measurements. 1

We consider the problem of wrapping around an object, of which two views are available, a reference surface and recovering the resulting parametric flow using direct computations (via spatio-temporal derivatives). The well known examples are affine flow models and eight-parameter flow models-both describing a flow field of a planar reference surface. We extend those classic flow models to deal with a quadric reference surface and work out the explicit parametric form of the flow field. As a result we derive a simple warping algorithm that maps between two views and leaves a residual flow proportional to the 3D deviation of the surface from a virtual quadric surface. The applications include image morphing, model building, image stabilization, and disparate view correspondence.

We consider a scene, containing many objects moving with constant velocity along straight line paths, seen from three reference viewpoints at three different times. The scene may even consist only of moving objects with no static features. We wish to create a new image sequence showing the scene from arbitrary viewing position and arbitrary time. We make use of a newly discovered tool, the "dual Htensor" that connects together three views of a coplanar configuration of (unlabeled) static and moving points. The newly synthesized images use constant velocity in the world to achieve realistic and physically correct images.

Introduces a family of 4/spl times/4/spl times/4 tensors, referred to as "join tensors" or Jtensors for short, which perform "3D to 3D" alignment between coordinate systems of sets of dynamic 3D points. 3D configurations of points are obtained by a 3D measuring device (such as a structured light or laser range sensor, or a stereo rig) at times t/sub 1/, t/sub 2/, t/sub 3/ from different viewing positions in addition to the motion of the sensor the points are also allowed to move in space; each point can move along an arbitrary straight-line path-we refer to this situation as "dynamic". The problem is to recover the motion of the sensor given the 3D correspondences of the points over time. We introduce Jtensors to capture the problem described above. Three observations P, P', P'' of a point measured at three time instants contribute a linear measurement to the Jtensor, regardless of whether the point has moved in space or has remained stationary while the sensor has changed position.

We consider the problem of wrapping around an object, of which two views are available, a reference surface and recovering the resulting parametric flow using direct computations (via spatio-temporal derivatives). The well known examples are affine flow models and B-parameter flow models - both describing a flow field of a planar reference surface. We extend those classic flow models to deal with a quadric reference surface and work out the explicit parametric form of the flow field. As a result we derive a simple warping algorithm that maps between two views and leaves a residual flow proportional to the 30 deviation of the surface from a virtual quadric surface. The applications include image morphing, model building, image stabilization, and disparate view correspondence.