Blind deconvolution problems are heavily ill-posed where the specific blurring kernel is not known. Recovering these images typically requires estimates of the kernel. In this paper, we present a method called Dr-Net, which does not require any such estimate and is further able to invert the effects of the blurring in blind image recovery tasks. These image recovery problems typically have two terms, the data fidelity term (for faithful reconstruction) and the image prior (for realistic looking reconstructions). We use the Douglas-Rachford iterations to solve this problem since it is a more generally applicable optimization procedure than methods such as the proximal gradient descent algorithm. Two proximal operators originate from these iterations, one from the data fidelity term and the second from the image prior. It is non-trivial to design a hand-crafted function to represent these proximal operators for the data fidelity and the image prior terms which would work with real-world image distributions. We therefore approximate both these proximal operators using deep networks. This provides a sound motivation for the final architecture for Dr-Net which we find outperforms the state-of-the-art on two mainstream blind deconvolution benchmarks. We also find that Dr-Net is one of the fastest algorithms according to wall-clock times while doing so.

Recently, impressive results have been achieved on many large-scale face recognition benchmarks, such as IJB-A Janus and Janus CS3. These datasets were designed to test robustness to nuisance transformations simultaneously such as pose, illumination, expression etc. We present a study paper, where we find that despite this goal in evaluation, there exists a significant frontal bias in yaw pose in these datasets. Therefore, high-performance on these recent datasets is misleading and does not reflect robustness to extreme pose in yaw. Moreover, many real-world applications only allow for a single frontal enrollment in a gallery (law enforcement, immigration etc.). As we show in our study, face recognition in this highly constrained setting with extreme pose variation in the probe images remains a highly challenging problem. Traditional approaches, performing well on datasets such as IJB-A Janus, perform much worse on older but highly controlled datasets such as CMU MPIE. To aid our study, we present a simple and practical method to handle pose variation in face recognition pipelines designed to deal with extremely off-angle faces. Our approach is to ignore the half of the face with any self-occlusion. This method allows our models to be highly robust to pose, and helps us achieve state-of-the-art results on several protocols using the CMU MPIE dataset as well as very accurate results on the CFP dataset, outperforming recent efforts using the same training data.

We motivate and present Ring loss, a simple and elegant feature normalization approach for deep networks designed to augment standard loss functions such as Softmax. We argue that deep feature normalization is an important aspect of supervised classification problems where we require the model to represent each class in a multi-class problem equally well. The direct approach to feature normalization through the hard normalization operation results in a non-convex formulation. Instead, Ring loss applies soft normalization, where it gradually learns to constrain the norm to the scaled unit circle while preserving convexity leading to more robust features. We apply Ring loss to large-scale face recognition problems and present results on LFW, the challenging protocols of IJB-A Janus, Janus CS3 (a superset of IJB-A Janus), Celebrity Frontal-Profile (CFP) and MegaFace with 1 million distractors. Ring loss outperforms strong baselines, matches state-of-the-art performance on IJB-A Janus and outperforms all other results on the challenging Janus CS3 thereby achieving state-of-the-art. We also outperform strong baselines in handling extremely low resolution face matching.

The study of representations invariant to common transformations of the data is important to learning. Most techniques have focused on local approximate invariance implemented within expensive optimization frameworks lacking explicit theoretical guarantees. In this paper, we study kernels that are invariant to a unitary group while having theoretical guarantees in addressing the important practical issue of unavailability of transformed versions of labelled data. A problem we call the Unlabeled Transformation Problem which is a special form of semi-supervised learning and one-shot learning. We present a theoretically motivated alternate approach to the invariant kernel SVM based on which we propose Max-Margin Invariant Features (MMIF) to solve this problem. As an illustration, we design an framework for face recognition and demonstrate the efficacy of our approach on a large scale semi-synthetic dataset with 153,000 images and a new challenging protocol on Labelled Faces in the Wild (LFW) while out-performing strong baselines.

We propose an explicitly discriminative and 'simple' approach to generate invariance to nuisance transformations modeled as unitary. In practice, the approach works well to handle non-unitary transformations as well. Our theoretical results extend the reach of a recent theory of invariance to discriminative and kernelized features based on unitary kernels. As a special case, a single common framework can be used to generate subject-specific pose-invariant features for face recognition and vice-versa for pose estimation. We show that our main proposed method (DIKF) can perform well under very challenging large-scale semisynthetic face matching and pose estimation protocols with unaligned faces using no landmarking whatsoever. We additionally benchmark on CMU MPIE and outperform previous work in almost all cases on off-angle face matching while we are on par with the previous state-of-the-art on the LFW unsupervised and image-restricted protocols, without any low-level image descriptors other than raw-pixels.

The search for new biometrics is never ending. In this work, we investigate the use of image based nasal features as a biometric. In many real-world recognition scenarios, partial occlusions on the face leave the nose region visible (e.g. sunglasses). Face recognition systems often fail or perform poorly in such settings. Furthermore, the nose region naturally contain more invariance to expression than features extracted from other parts of the face. In this study, we extract discriminative nasal features using Kernel Class-Dependence Feature Analysis (KCFA) based on Optimal Trade-off Synthetic Discriminant Function (OTSDF) filters. We evaluate this technique on the FRGC ver2.0 database and the AR Face database, training and testing exclusively on nasal features and have compared the results to the full face recognition using KCFA features. We find that the between-subject discriminability in nasal features is comparable to that found in facial features. This shows that nose biometrics have a potential to support and boost biometric identification, that has largely been under utilized. Moreover, our extracted KCFA nose features have significantly outperformed the PittPatt face matcher which works with the original JPEG images on the AR facial occlusion database. This shows that nose biometrics can be used as a stand-alone biometric trait when the subjects are under occlusions.

Many optimization problems are multi-modal. In certain cases, we are interested in finding multiple locally optimal solutions rather than just a single optimum as is computed by traditional genetic algorithms (GAs). Several niching techniques have been developed that seek to find multiple such local optima. These techniques, which include sharing and crowding, are clearly powerful and useful. But they do not explicitly let the user control the number of local optima being computed, which we believe to be an important capability. In this paper, we develop a method that provides, as an input parameter to niching, the desired number of local optima. Our method integrates techniques from feedback control, includes a sensor based on clustering, and utilizes a scaling parameter in Generalized Crowding to control the number of niches being explored. The resulting Feedback Control GA (FCGA) is tested in several experiments and found to perform well compared to previous approaches. Overall, the integration of feedback control and Generalized Crowding is shown to effectively guide the search for multiple local optima in a more controlled fashion. We believe this novel capability has the potential to impact future applications as well as other evolutionary algorithms.