Knowing when a user picks up an object plays a vital role in many context-aware applications. For example, tracking water consumption, counting calories consumed, or reminding you to bring your keys are all context-centered scenarios involving picking up objects. In this project, we propose Contextra, a wrist-worn system that uses sensor fusion to recognize when a user grasps objects. Sensor fusion allows all parts of the grasp to be sensed in ways single channels cannot alone. In our wristband, we fuse EMG and IMU data with video captured from three low-power IR cameras. These cameras maintain privacy by using an active-illumination technique to only capture features close to the sensors. Beyond grasps alone, we see Contextra as playing a foundational role in providing continuous awareness of context triggers to extend the functionality of existing AI devices that cannot run continuously due to power and privacy concerns.

Smart rings for subtle, reliable finger input offer an attractive path for ubiquitous interaction with wearable computing platforms. However, compared to ordinary rings worn for cultural or fashion reasons, smart rings are much bulkier and less comfortable, largely due to the space required for a battery, which also limits the space available for sensors. This paper presents picoRing, a flexible sensing architecture that enables a variety of battery-free smart rings paired with a wristband. By inductively connecting a wristband-based sensitive reader coil with a ring-based fully-passive sensor coil, picoRing enables the wristband to stably detect the passive response from the ring via a weak inductive coupling. We demonstrate four different rings that support thumb-to-finger interactions like pressing, sliding, or scrolling. When users perform these interactions, the corresponding ring converts each input into a unique passive response through a network of passive switches. Combining the coil-based sensitive readout with the fully-passive ring design enables a tiny ring that weighs as little as 1.5 g and achieves a 13 cm stable readout despite finger bending, and proximity to metal.

We present ElectroRing, a wearable ring-based input device that reliably detects both onset and release of a subtle finger pinch, and more generally, contact of the fingertip with the user’s skin. ElectroRing addresses a common problem in ubiquitous touch interfaces, where subtle touch gestures with little movement or force are not detected by a wearable camera or IMU. ElectroRing’s active electrical sensing approach provides a step-function-like change in the raw signal, for both touch and release events, which can be easily detected using only basic signal processing techniques. Notably, ElectroRing requires no second point of instrumentation, but only the ring itself, which sets it apart from existing electrical touch detection methods. We built three demo applications to highlight the effectiveness of our approach when combined with a simple IMU-based 2D tracking system.

We introduce RotoWrist, an infrared (IR) light based solution for continuously and reliably tracking 2-degree-of-freedom (DoF) relative angle of the wrist with respect to the forearm using a wristband. The tracking system consists of eight time-of-flight (ToF) IR light modules distributed around a wristband. We developed a computationally simple tracking approach to reconstruct the orientation of the wrist without any runtime training, ensuring user independence. An evaluation study demonstrated that RotoWrist achieves a cross-user median tracking error of 5.9° in flexion/extension and 6.8° in radial and ulnar deviation with no calibration required as measured with optical ground truth. We further demonstrate the performance of RotoWrist for a pointing task and compare it against ground truth tracking.

Contemporary AR/VR systems use in-air gestures or handheld controllers for interactivity. This overlooks the skin as a convenient surface for tactile, touch-driven interactions, which are generally more accurate and comfortable than free space interactions. In response, we developed ActiTouch, a new electrical method that enables precise on-skin touch segmentation by using the body as an RF waveguide. We combine this method with computer vision, enabling a system with both high tracking precision and robust touch detection. Our system requires no cumbersome instrumentation of the fingers or hands, requiring only a single wristband (e.g., smartwatch) and sensors integrated into an AR/VR headset. We quantify the accuracy of our approach through a user study and demonstrate how it can enable touchscreen-like interactions on the skin.

Long videos can be played much faster than real-time by recording only one frame per second or by dropping all but one frame each second, i.e., by creating a timelapse. Unstable hand-held moving videos can be stabilized with a number of recently described methods. Unfortunately, creating a stabilized timelapse, or hyperlapse , cannot be achieved through a simple combination of these two methods. Two hyperlapse methods have been previously demonstrated: one with high computational complexity and one requiring special sensors. We present an algorithm for creating hyperlapse videos that can handle significant high-frequency camera motion and runs in real-time on HD video. Our approach does not require sensor data, thus can be run on videos captured on any camera. We optimally select frames from the input video that best match a desired target speed-up while also resulting in the smoothest possible camera motion. We evaluate our approach using several input videos from a range of cameras and compare these results to existing methods.

We present LightRing, a wearable sensor in a ring form factor that senses the 2d location of a fingertip on any surface, independent of orientation or material. The device consists of an infrared proximity sensor for measuring finger flexion and a 1-axis gyroscope for measuring finger rotation. Notably, LightRing tracks subtle fingertip movements from the finger base without requiring instrumentation of other body parts or the environment. This keeps the normal hand function intact and allows for a socially acceptable appearance. We evaluate LightRing in a 2d pointing experiment in two scenarios: on a desk while sitting down, and on the leg while standing. Our results indicate that the device has potential to enable a variety of rich mobile input scenarios.

We present a method for composing handwritten messages on a small touchscreen device. A word is entered by drawing overlapped, screen sized letters on top of each other. The system does not require gestures or timeouts to delimit characters within a word - it automatically segments the overlapping strokes and renders the message in real-time as the user is writing. The auto-segmentation algorithm was designed for practicality; it is extremely simple, requires only public domain data for training, and runs very fast on low-power devices. Drawings may also be included with the text. Experimental data indicates the effectiveness of our system, even for novice users.

We present a data-driven method to predict the quality of an image completion method. Our method is based on the state-of-the-art non-parametric framework of Wexleret al. [2007]. It uses automatically derived search space constraints for patch source regions, which lead to improved texture synthesis and semantically more plausible results. These constraints also facilitate performance prediction by allowing us to correlate output quality against features of possible regions used for synthesis. We use our algorithm to first crop and then complete stitched panoramas. Our predictive ability is used to find an optimal crop shapebeforethe completion is computed, potentially saving significant amounts of computation. Our optimized crop includes as much of the original panorama as possible while avoiding regions that can be less successfully filled in. Our predictor can also be applied for hole filling in the interior of images. In addition to extensive comparative results, we ran several user studies validating our predictive feature, good relative quality of our results against those of other state-of-the-art algorithms, and our automatic cropping algorithm.

This paper presents a fully automated algorithm for reconstructing a textured 3D model of a face from a single photograph or a raw video stream. The algorithm is based on a combination of Support Vector Machines (SVMs) and a Morphable Model of 3D faces. After SVM face detection, individual facial features are detected using a novel regression- and classification-based approach, and probabilistically plausible configurations of features are selected to produce a list of candidates for several facial feature positions. In the next step, the configurations of feature points are evaluated using a novel criterion that is based on a Morphable Model and a combination of linear projections. To make the algorithm robust with respect to head orientation, this process is iterated while the estimate of pose is refined. Finally, the feature points initialize a model-fitting procedure of the Morphable Model. The result is a highresolution 3D surface model.

We present a new approach to personalized handwriting recognition. The problem, also known as writer adaptation, consists of converting a generic (user-independent) recognizer into a personalized (user-dependent) one, which has an improved recognition rate for a particular user. The adaptation step usually involves user-specific samples, which leads to the fundamental question of how to fuse this new information with that captured by the generic recognizer. We propose adapting the recognizer by minimizing a regularized risk functional (a modified SVM) where the prior knowledge from the generic recognizer enters through a modified regularization term. The result is a simple personalization framework with very good practical properties. Experiments on a 100 class real-world data set show that the number of errors can be reduced by over 40% with as few as five user samples per character.

This paper addresses the bottom-up influence of local image information on human eye movements. Most existing computational models use a set of biologically plausible linear filters, e.g., Gabor or Difference-of-Gaussians filters as a front-end, the outputs of which are nonlinearly combined into a real number that indicates visual saliency. Unfortunately, this requires many design parameters such as the number, type, and size of the front-end filters, as well as the choice of nonlinearities, weighting and normalization schemes etc., for which biological plausibility cannot always be justified. As a result, these parameters have to be chosen in a more or less ad hoc way. Here, we propose to learn a visual saliency model directly from human eye movement data. The model is rather simplistic and essentially parameter-free, and therefore contrasts recent developments in the field that usually aim at higher prediction rates at the cost of additional parameters and increasing model complexity. Experimental results show that--despite the lack of any biological prior knowledge--our model performs comparably to existing approaches, and in fact learns image features that resemble findings from several previous studies. In particular, its maximally excitatory stimuli have center-surround structure, similar to receptive fields in the early human visual system.

This paper proposes a method for computing fast approximations to sup- port vector decision functions in the field of object detection. In the present approach we are building on an existing algorithm where the set of support vectors is replaced by a smaller, so-called reduced set of syn- thesized input space points. In contrast to the existing method that finds the reduced set via unconstrained optimization, we impose a structural constraint on the synthetic points such that the resulting approximations can be evaluated via separable filters. For applications that require scan- ning large images, this decreases the computational complexity by a sig- nificant amount. Experimental results show that in face detection, rank deficient approximations are 4 to 6 times faster than unconstrained re- duced set systems. 1 Introduction It has been shown that support vector machines (SVMs) provide state-of-the-art accuracies in object detection. In time-critical applications, however, they are of limited use due to their computationally expensive decision functions. In particular, the time complexity of an SVM classification operation is characterized by two parameters. First, it is linear in the number of support vectors (SVs). Second, it scales with the number of operations needed for computing the similarity between an SV and the input, i.e. the complexity of the kernel function. When classifying image patches of size h w using plain gray value features, the decision function requires an h w dimensional dot product for each SV. As the patch size increases, these computations become extremely expensive. As an example, the evaluation of a single 20 20 patch on a 320 240 image at 25 frames per second already requires 660 million operations per second. In the past, research towards speeding up kernel expansions has focused exclusively on the first issue, i.e. on how to reduce the number of expansion points (SVs) [1, 2]. In [2], Burges introduced a method that, for a given SVM, creates a set of so-called reduced set vectors (RSVs) that approximate the decision function. This approach has been successfully ap- plied in the image classification domain -- speedups on the order of 10 to 30 have been re- ported [2, 3, 4] while the full accuracy was retained. Additionally, for strongly unbalanced classification problems such as face detection, the average number of RSV evaluations can be further reduced using cascaded classifiers [5, 6, 7]. Unfortunately, the above example illustrates that even with as few as three RSVs on average (as in [5]), such systems are not competitive for time-critical applications. The present work focuses on the second issue, i.e. the high computational cost of the kernel evaluations. While this could be remedied by switching to a sparser image representation (e.g. a wavelet basis), one could argue that in connection with SVMs, not only are plain gray values straightforward to use, but they have shown to outperform Haar wavelets and gradients in the face detection domain [8]. Alternatively, in [9], the authors suggest to com- pute the costly correlations in the frequency domain. In this paper, we develop a method that combines the simplicity of gray value correlations with the speed advantage of more sophisticated image representations. To this end, we borrow an idea from image processing: by constraining the RSVs to have a special structure, they can be evaluated via separable convolutions. This works for most standard kernels (e.g. linear, polynomial, Gaussian and sigmoid) and decreases the average computational complexity of the RSV evaluations from O(h w) to O(r (h + w)), where r is a small number that allows the user to balance be- tween speed and accuracy. To evaluate our approach, we examine the performance of these approximations on the MIT+CMU face detection database (used in [10, 8, 5, 6]). 2 Burges' method for reduced set approximations The present section briefly describes Burges' reduced set method [2] on which our work is based. For reasons that will become clear below, h w image patches are written as h w matrices (denoted by bold capital letters) whose entries are the respective pixel intensities. In this paper, we refer to this as the image-matrix notation. Assume that an SVM has been successfully trained on the problem at hand. Let {X1, . . . Xm} denote the set of SVs, {1, . . . m} the corresponding coefficients, k(, ) the kernel function and b the bias of the SVM solution. The decision rule for a test pattern X reads m f (X) = sgn yiik(Xi, X) + b . (1) i=1 In SVMs, the decision surface induced by f corresponds to a hyperplane in the reproducing kernel Hilbert space (RKHS) associated with k. The corresponding normal m = yiik(Xi, ) (2) i=1 can be approximated using a smaller, so-called reduced set (RS) {Z1, . . . Zm } of size m < m, i.e. an approximation to of the form m = ik(Zi, ). (3) i=1 This speeds up the decision process by a factor of m/m . To find such , we fix a desired set size m and solve min - 2RKHS (4) for i and Zi. Here, RKHS denotes the Euclidian norm in the RKHS. The resulting RS decision function f is then given by m f (X) = sgn ik(Zi,X)+b i=1 . (5) In practice, i, Zi are found using a gradient based optimization technique. Details can be found in [2]. 3 From separable filters to rank deficient reduced sets We now describe the concept of separable filters in image processing and show how this idea extends to a broader class of linear filters and to a special class of nonlinear filters, namely those used by SVM decision functions. Using the image-matrix notation, it will become clear that the separability property boils down to a matrix rank constraint. 3.1 Linear separable filters Applying a linear filter to an image amounts to a two-dimensional convolution of the image with the impulse response of the filter. In particular, if I is the input image, H the impulse response, i.e. the filter mask, and J the output image, then J = I H. (6) If H has size h w, the convolution requires O(h w) operations for each output pixel. However, in special cases where H can be decomposed into two column vectors a and b, such that H = ab (7) holds, we can rewrite (6) as J = [I a] b , (8) since the convolution is associative and in this case, ab = a b . This splits the original problem (6) into two convolution operations with masks of size h1 and 1w, respectively. As a result, if a linear filter is separable in the sense of equation (7), the computational complexity of the filtering operation can be reduced from O(h w) to O(h + w) per pixel by computing (8) instead of (6). 3.2 Linear rank deficient filters In view of (7) being equivalent to rank(H) 1, we now generalize the above concept to linear filters with low rank impulse responses. Consider the singular value decomposition (SVD) of the h w matrix H, H = USV , (9) and recall that U and V are orthogonal matrices of size h h and w w, respectively, whereas S is diagonal (the diagonal entries are the singular values) and has size h w. Now let r = rank(H). Due to rank(S) = rank(H), we may write H as a sum of r rank one matrices r H = s u v i i i (10) i=1 where si denotes the ith singular value of H and ui, vi are the iths columns of U and V (i.e. the ith singular vectors of the matrix H), respectively. As a result, the correspond- ing linear filter can be evaluated (analogously to (8)) as the weighted sum of r separable convolutions r J = si [I ui] vi (11) i=1 and the computational complexity drops from O(h w) to O(r (h + w)) per output pixel. Not surprisingly, the speed benefit depends on r, which can be seen to measure the structural complexity1 of H. For square matrices (w = h) for instance, (11) does not give any speedup compared to (6) if r > w/2. 1In other words, the flatter the spectrum of HH , the less benefit can be expected from (11). 3.3 Nonlinear rank deficient filters and reduced sets Due to the fact that in 2D, correlation is identical with convolution if the filter mask is rotated by 180 degrees (and vice versa), we can apply the above idea to any image filter f (X) = g(c(H, X)) where g is an arbitrary nonlinear function and c(H, X) denotes the correlation between images patches X and H (both of size h w). In SVMs this amounts to using a kernel of the form k(H, X) = g(c(H, X)). (12) If H has rank r, we may split the kernel evaluation into r separable correlations plus a scalar nonlinearity. As a result, if the RSVs in a kernel expansion such as (5) satisfy this constraint, the average computational complexity decreases from O(m h w) to O(m r (h + w)) per output pixel. This concept works for many off-the-shelf kernels used in SVMs. While linear, polynomial and sigmoid kernels are defined as functions of input space dot products and therefore immediately satisfy equation (12), the idea applies to kernels based on the Euclidian distance as well. For instance, the Gaussian kernel reads k(H, X) = exp((c(X, X) - 2c(H, X) + c(H, H))). (13) Here, the middle term is the correlation which we are going to evaluate via separable filters. The first term is independent of the SVs -- it can be efficiently pre-computed and stored in a separate image. The last term is merely a constant scalar independent of the image data. Finally, note that these kernels are usually defined on vectors. Nevertheless, we can use our image-matrix notation due to the fact that the squared Euclidian distance between two vectors of gray values x and z may be written as x - z 2 = X - Z 2 , (14) F whereas the dot product amounts to 1 x z = X 2 + Z 2 - X - Z 2 , (15) 2 F F F where X and Z are the corresponding image patches and F is the Frobenius norm for matrices. 4 Finding rank deficient reduced sets In our approach we consider a special class of the approximations given by (3), namely those where the RSVs can be evaluated efficiently via separable correlations. In order to obtain such approximations, we use a constrained version of Burges' method. In particular, we restrict the RSV search space to the manifold spanned by all image patches that -- viewed as matrices -- have a fixed, small rank r (which is to be chosen a priori by the user). To this end, the Zi in equation (3) are replaced by their singular value decompositions Z S V i Ui i i . (16) The rank constraint can then be imposed by allowing only the first r diagonal elements of Si to be non-zero. Note that this boils down to using an approximation of the form m = S V r ik(Ui,r i,r i,r , ) (17) i=1 with Si,r being r r (diagonal) and Ui,r, Vi,r being h r, w r (orthogonal2) matrices, respectively. Analogously to (4) we fix m and r and find Si,r, Ui,r, Vi,r and i that mini- mize the approximation error = - 2 . r The minimization problem is solved via RKHS 2In this paper we call a non-square matrix orthogonal if its columns are pairwise orthogonal and have unit length. gradient decent. Note that when computing gradients, the image-matrix notation (together with (14) or (15), and the equality X 2 = tr(XX )) allows a straightforward computa- F tion of the kernel derivatives w.r.t. the components of the decomposed RSV image patches, i.e. the row, column and scale information in Vi,r, Ui,r and Si,r, respectively. However, while the update rules for i and Si,r follow immediately from the respective derivatives, care must be taken in order to keep Ui,r and Vi,r orthogonal during optimization. This can be achieved through re-orthogonalization of these matrices after each gradient step. In our current implementation, however, we perform those updates subject to the so-called Stiefel constraints [11]. Intuitively, this amounts to rotating (rather than translating) the columns of Ui,r and Vi,r, which ensures that the resulting matrices are still orthogonal, i.e. lie on the Stiefel manifold. Let S(h, r) be the manifold of orthogonal h r matrices, the (h, r)-Stiefel manifold. Further, let U denote an orthogonal basis for the orthogo- i,r nal complement of the subspace spanned by the columns of Ui,r. Now, given the 'free' gradient G = /Ui,r we compute the 'constrained' gradient ^ G = G - U G U i,r i,r , (18) which is the projection of G onto the tangent space of S(h, r) at Ui,r. The desired rotation is then given [11] by the (matrix) exponential of the h h skew-symmetric matrix ^ G U ) A = t i,r -( ^ G U i,r ^ (19) G U 0 i,r where t is a user-defined step size parameter. For details, see [11]. A Matlab library is available at [12]. 5 Experiments This section shows the results of two experiments. The first part illustrates the behavior of rank deficient approximations for a face detection SVM in terms of the convergence rate and classification accuracy for different values of r. In the second part, we show how an actual face detection system, similar to that presented in [5], can be sped up using rank deficient RSVs. In both experiments we used the same training and validation set. It consisted of 19 19 gray level image patches containing 16081 manually collected faces (3194 of them kindly provided by Sami Romdhani) and 42972 non-faces automatically collected from a set of 206 background scenes. Each patch was normalized to zero mean and unit variance. The set was split into a training set (13331 faces and 35827 non-faces) and a validation set (2687 faces and 7145 non-faces). We trained a 1-norm soft margin SVM on the training set using a Gaussian kernel with = 10. The regularization constant C was set to 1. The resulting decision function (1) achieved a hit rate of 97.3% at 1.0% false positives on the validation set using m = 6910 SVs. This solution served as the approximation target (see equation (2)) during the experiments described below. 5.1 Rank deficient faces In order to see how m and r affect the accuracy of our approximations, we compute rank deficient reduced sets for m = 1 . . . 32 and r = 1 . . . 3 (the left array in Figure 1 illustrates the actual appearance of rank deficient RSVs for the m = 6 case). Accuracy of the re- sulting decision functions is measured in ROC score (the area under the ROC curve) on the validation set. For the full SVM, this amounts to 0.99. The results for our approximations are depicted in Figure 2. As expected, we need a larger number of rank deficient RSVs than unconstrained RSVs to obtain similar classification accuracies, especially for small r. Nevertheless, the experiment points out two advantages of our method. First, a rank as m' = 6 m' = 1 = + + + ... r=full r=full r=3 = + + r=3 r=2 = + r=2 r=1 r=1 = Figure 1: Rank deficient faces. The left array shows the RSVs (Zi) of the unconstrained (top row) and constrained (r decreases from 3 to 1 down the remaining rows) approxi- mations for m = 6. Interestingly, the r = 3 RSVs are already able to capture face-like structures. This supports the fact that the classification accuracy for r = 3 is similar to that of the unconstrained approximations (cf. Figure 2, left plot). The right array shows the m = 1 RSVs (r = full, 3, 2, 1, top to bottom row) and their decomposition into rank one matrices according to (10). For the unconstrained RSV (first row) it shows an approximate (truncated) expansion based on the three leading singular vectors. While for r = 3 the decomposition is indeed similar to the truncated SVD, note how this similarity decreases for r = 2, 1. This illustrates that the approach is clearly different from simply finding un- constrained RSVs and then imposing the rank constraint via SVD (in fact, the norm (4) is smaller for the r = 1 RSV than for the leading singular vector of the r = full RSV). low as three seems already sufficient for our face detection SVM in the sense that for equal sizes m there is no significant loss in accuracy compared to the unconstrained approxima- tion (at least for m > 2). The associated speed benefit over unconstrained RSVs is shown in the right plot of Figure 2: the rank three approximations achieve accuracies similar to the unconstrained functions, while the number of operations reduces to less than a third. Second, while for unconstrained RSVs there is no solution with a number of operations smaller than h w = 361 (in the right plot, this is the region beyond the left end of the solid line), there exist rank deficient functions which are not only much faster than this, but yield considerably higher accuracies. This property will be exploited in the next experiment. 5.2 A cascade-based face detection system In this experiment we built a cascade-based face detection system similar to [5, 6], i.e. a cascade of RSV approximations of increasing size m . As the benefit of a cascaded classifier heavily depends on the speed of the first classifier which has to be evaluated on the whole image [5, 6], our system uses a rank deficient approximation as the first stage. Based on the previous experiment, we chose the m = 3, r = 1 classifier. Note that this function yields an ROC score of 0.9 using 114 multiply-adds, whereas the simplest possible unconstrained approximation m = 1, r = full needs 361 multiply-adds to achieve a ROC score of only 0.83 (cf. Figure 2). In particular, if the threshold of the first stage is set to yield a hit rate of 95% on the validation set, scanning the MIT+CMU set (130 images, 507 faces) with m = 3, r = 1 discards 91.5% of the false positives, whereas the m = 1, r = full can only reject 70.2%. At the same time, when scanning a 320 240 image3, the three separable convolutions plus nonlinearity require 55ms, whereas the single, full kernel evaluation takes 208ms on a Pentium 4 with 2.8 GHz. Moreover, for the unconstrained 3For multi-scale processing the detectors are evaluated on an image pyramid with 12 different scales using a scale decay of 0.75. This amounts to scanning 140158 patches for a 320 240 image. 1 1 0.9 0.9 0.8 0.8 ROC score ROC score r=1 r=1 0.7 r=2 0.7 r=2 r=3 r=3 r=full r=full 0.6 0.6 0 1 2 3 4 10 10 10 10 10 #RSVs (m') #operations (m'r(h+w)) Figure 2: Effect of the rank parameter r on classification accuracies. The left plots shows the ROC score of the rank deficient RSV approximations (cf. Section 4) for varying set sizes (m = 1 . . . 32, on a logarithmic scale) and ranks (r = 1 . . . 3). Additionally, the solid line shows the accuracy of the RSVs without rank constraint (cf. Section 2), here denoted by r = full. The right plot shows the same four curves, but plotted against the number of operations needed for the evaluation of the corresponding decision function when scanning large images (i.e. m r (h + w) with h = w = 19), also on a logarithmic scale. Figure 3: A sample output from our demonstra- tion system (running at 14 frames per second). In this implementation, we reduced the number of false positives by adjusting the threshold of the final classifier. Although this reduces the number of detections as well, the results are still satisfactory. This is probably due to the fact that the MIT+CMU set contains several images of very low quality that are not likely to occur in our setting, using a good USB camera. cascade to catch up in terms of accuracy, the (at least) m = 2, r = full classifier (also with an ROC score of roughly 0.9) should be applied afterwards, requiring another 0.3 2 208 ms 125ms. The subsequent stages of our system consist of unconstrained RSV approximations of size m = 4, 8, 16, 32, respectively. These sizes were chosen such that the number of false positives roughly halves after each stage, while the number of correct detections remains close to 95% on the validation set (with the decision thresholds adjusted accordingly). To eliminate redundant detections, we combine overlapping detections via averaging of posi- tion and size if they are closer than 0.15 times the estimated patch size. This system yields 93.1% correct detections and 0.034% false positives on the MIT+CMU set. The current system was incorporated into a demo application (Figure 3). For optimal performance, we re-compiled our system using the Intel compiler (ICC). The application now classifies a 320x240 image within 54ms (vs. 238ms with full rank RSVs only) on a 2.8 GHz PC. To further reduce the number of false positives, additional bootstrapped (as in [5]) stages need to be added to the cascade. Note that this will not significantly affect the speed of our sys- tem (currently 14 frames per second) since 0.034% false positives amounts to merely 47 patches to be processed by subsequent classifiers.