Springer Science and Business Media LLC

The prediction sum of squares is a useful statistic for comparing different models. It is based on the principle of leave-one-out or ordinary cross-validation, whereby every measurement is considered in turn as a test set, for the model parameters trained on all but the held out measurement. As for linear least squares problems, there is a simple well-known non-iterative formula to compute the prediction sum of squares without having to refit the model as many times as the number of measurements. We extend this formula to cases where the problem has multiple parameter or measurement sets. We report experimental results on the fitting of a warp between two images, for which the number of deformation centres is automatically selected, based on one of the proposed non-iterative formulae.

In this paper, we propose a method for initial camera pose estimation from just a single image which is robust to viewing conditions and does not require a detailed model of the scene. This method meets the growing need of easy deployment of robotics or augmented reality applications in any environments, especially those for which no accurate 3D model nor huge amount of ground truth data are available. It exploits the ability of deep learning techniques to reliably detect objects regardless of viewing conditions. Previous works have also shown that abstracting the geometry of a scene of objects by an ellipsoid cloud allows to compute the camera pose accurately enough for various application needs. Though promising, these approaches use the ellipses fitted to the detection bounding boxes as an approximation of the imaged objects. In this paper, we go one step further and propose a learning-based method which detects improved elliptic approximations of objects which are coherent with the 3D ellipsoids in terms of perspective projection. Experiments prove that the accuracy of the computed pose significantly increases thanks to our method. This is achieved with very little effort in terms of training data acquisition—a few hundred calibrated images of which only three need manual object annotation. Code and models are released at https://gitlab.inria.fr/tangram/3d-aware-ellipses-for-visual-localization .

Semantic segmentation with limited annotations, such as weakly supervised semantic segmentation (WSSS) and semi-supervised semantic segmentation (SSSS), is a challenging task that has attracted much attention recently. Most leading WSSS methods employ a sophisticated multi-stage training strategy to estimate pseudo-labels as precise as possible, but they suffer from high model complexity. In contrast, there exists another research line that trains a single network with image-level labels in one training cycle. However, such a single-stage strategy often performs poorly because of the compounding effect caused by inaccurate pseudo-label estimation. To address this issue, this paper presents a Self-supervised Low-Rank Network (SLRNet) for single-stage WSSS and SSSS. The SLRNet uses cross-view self-supervision, that is, it simultaneously predicts several complementary attentive LR representations from different views of an image to learn precise pseudo-labels. Specifically, we reformulate the LR representation learning as a collective matrix factorization problem and optimize it jointly with the network learning in an end-to-end manner. The resulting LR representation deprecates noisy information while capturing stable semantics across different views, making it robust to the input variations, thereby reducing overfitting to self-supervision errors. The SLRNet can provide a unified single-stage framework for various label-efficient semantic segmentation settings: (1) WSSS with image-level labeled data, (2) SSSS with a few pixel-level labeled data, and (3) SSSS with a few pixel-level labeled data and many image-level labeled data. Extensive experiments on the Pascal VOC 2012, COCO, and L2ID datasets demonstrate that our SLRNet outperforms both state-of-the-art WSSS and SSSS methods with a variety of different settings, proving its good generalizability and efficacy.

The computational problems in segmenting range data into surface patches based on the invariant surface properties, i.e., mean curvature H and Gaussian curvature K, are investigated. The goal is to obtain reliable HK surface maps. Two commonly encountered problems are: firstly the noise effect in computing derivative estimates, and secondly the smoothing across discontinuities. Here, the segmentation is formulated as finding minimization solutions of energy functionals involving discontinuities. A two-stage approach to the goal is presented: stage (1) from a range image to curvature images and stage (2) from the curvature images to the HK maps. In both stages, solutions are found through minimizing energy functionals that measure the degree of bias of a solution from two constraints: the closeness of the solution to the data, and the smoothness of the solution controlled by predetermined discontinuities. Propagation across discontinuities is prevented during minimization, which preserves the original surface shapes. Experimental results are given for a variety of test images.

This paper comprehensively surveys the development of face hallucination (FH), including both face super-resolution and face sketch-photo synthesis techniques. Indeed, these two techniques share the same objective of inferring a target face image (e.g. high-resolution face image, face sketch and face photo) from a corresponding source input (e.g. low-resolution face image, face photo and face sketch). Considering the critical role of image interpretation in modern intelligent systems for authentication, surveillance, law enforcement, security control, and entertainment, FH has attracted growing attention in recent years. Existing FH methods can be grouped into four categories: Bayesian inference approaches, subspace learning approaches, a combination of Bayesian inference and subspace learning approaches, and sparse representation-based approaches. In spite of achieving a certain level of development, FH is limited in its success by complex application conditions such as variant illuminations, poses, or views. This paper provides a holistic understanding and deep insight into FH, and presents a comparative analysis of representative methods and promising future directions.

Comprehending different categories of facial expressions plays a great role in the design of computational model analyzing human perceived and affective state. Authoritative studies have revealed that facial expressions in human daily life are in multiple or co-occurring mental states. However, due to the lack of valid datasets, most previous studies are still restricted to basic emotions with single label. In this paper, we present a novel multi-label facial expression database, RAF-ML, along with a new deep learning algorithm, to address this problem. Specifically, a crowdsourcing annotation of 1.2 million labels from 315 participants was implemented to identify the multi-label expressions collected from social network, then EM algorithm was designed to filter out unreliable labels. For all we know, RAF-ML is the first database in the wild that provides with crowdsourced cognition for multi-label expressions. Focusing on the ambiguity and continuity of blended expressions, we propose a new deep manifold learning network, called Deep Bi-Manifold CNN, to learn the discriminative feature for multi-label expressions by jointly preserving the local affinity of deep features and the manifold structures of emotion labels. Furthermore, a deep domain adaption method is leveraged to extend the deep manifold features learned from RAF-ML to other expression databases under various imaging conditions and cultures. Extensive experiments on the RAF-ML and other diverse databases (JAFFE, CK $$+$$ , SFEW and MMI) show that the deep manifold feature is not only superior in multi-label expression recognition in the wild, but also captures the elemental and generic components that are effective for a wide range of expression recognition tasks.

The goal of single image deblurring is to recover both a latent clear image and an underlying blur kernel from one input blurred image. Recent methods focus on exploiting natural image priors or additional image observations for deblurring, but pay less attention to the influence of image structure on estimating blur kernels. What is the useful image structure and how can one select good regions for deblurring? We formulate the problem of learning good regions for deblurring within the conditional random field framework. To better compare blur kernels, we develop an effective similarity metric for labeling training samples. The learned model is able to predict good regions from an input blurred image for deblurring without user guidance. Qualitative and quantitative evaluations demonstrate that good regions can be selected by the proposed algorithms for effective single image deblurring.

Joint image filters are used to transfer structural details from a guidance picture used as a prior to a target image, in tasks such as enhancing spatial resolution and suppressing noise. Previous methods based on convolutional neural networks (CNNs) combine nonlinear activations of spatially-invariant kernels to estimate structural details and regress the filtering result. In this paper, we instead learn explicitly sparse and spatially-variant kernels. We propose a CNN architecture and its efficient implementation, called the deformable kernel network (DKN), that outputs sets of neighbors and the corresponding weights adaptively for each pixel. The filtering result is then computed as a weighted average. We also propose a fast version of DKN that runs about seventeen times faster for an image of size $$640 \times 480$$ . We demonstrate the effectiveness and flexibility of our models on the tasks of depth map upsampling, saliency map upsampling, cross-modality image restoration, texture removal, and semantic segmentation. In particular, we show that the weighted averaging process with sparsely sampled $$3 \times 3$$ kernels outperforms the state of the art by a significant margin in all cases.

This paper proposes a learning-based quality evaluation framework for inpainted results that does not require any subjectively annotated training data. Image inpainting, which removes and restores unwanted regions in images, is widely acknowledged as a task whose results are quite difficult to evaluate objectively. Thus, existing learning-based image quality assessment (IQA) methods for inpainting require subjectively annotated data for training. However, subjective annotation requires huge cost and subjects’ judgment occasionally differs from person to person in accordance with the judgment criteria. To overcome these difficulties, the proposed framework generates and uses simulated failure results of inpainted images whose subjective qualities are controlled as the training data. We also propose a masking method for generating training data towards fully automated training data generation. These approaches make it possible to successfully estimate better inpainted images, even though the task is quite subjective. To demonstrate the effectiveness of our approach, we test our algorithm with various datasets and show it outperforms existing IQA methods for inpainting.