Digital artists often improve the aesthetic quality of digital photographs through manual retouching. Beyond global adjustments, professional image editing programs provide local adjustment tools operating on specific parts of an image. Options include parametric (graduated, radial filters) and unconstrained brush tools. These highly expressive tools enable a diverse set of local image enhancements. However, their use can be time consuming, and requires artistic capability. State-of-the-art automated image enhancement approaches typically focus on learning pixel-level or global enhancements. The former can be noisy and lack interpretability, while the latter can fail to capture fine-grained adjustments. In this paper, we introduce a novel approach to automatically enhance images using learned spatially local filters of three different types (Elliptical Filter, Graduated Filter, Polynomial Filter). We introduce a deep neural network, dubbed Deep Local Parametric Filters (DeepLPF), which regresses the parameters of these spatially localized filters that are then automatically applied to enhance the image. DeepLPF provides a natural form of model regularization and enables interpretable, intuitive adjustments that lead to visually pleasing results. We report on multiple benchmarks and show that DeepLPF produces state-of-the-art performance on two variants of the MIT-Adobe 5k dataset, often using a fraction of the parameters required for competing methods.

We present a novel approach to adjust global image properties such as colour, saturation, and luminance using human-interpretable image enhancement curves, inspired by the Photoshop curves tool. Our method, dubbed neural CURve Layers (CURL), is designed as a multi-colour space neural retouching block trained jointly in three different colour spaces (HSV, CIELab, RGB) guided by a novel multi-colour space loss. The curves are fully differentiable and are trained end-to-end for different computer vision problems including photo enhancement (RGB-to-RGB) and as part of the image signal processing pipeline for image formation (RAW-to-RGB). To demonstrate the effectiveness of CURL we combine this global image transformation block with a pixel-level (local) image multi-scale encoder-decoder backbone network. In an extensive experimental evaluation we show that CURL produces state-of-the-art image quality versus recently proposed deep learning approaches in both objective and perceptual metrics, setting new state-of-the-art performance on multiple public datasets. Our code is publicly available at: https://github.com/sjmoran/CURL.

In this paper we focus on improving the effectiveness of hashing-based approximate nearest neighbour search. Generating similarity preserving hashcodes for images has been shown to be an effective and efficient method for searching through large datasets. Hashcode generation generally involves two steps: bucketing the input feature space with a set of hyperplanes, followed by quantising the projection of the data-points onto the normal vectors to those hyperplanes. This procedure results in the makeup of the hashcodes depending on the positions of the data-points with respect to the hyperplanes in the feature space, allowing a degree of locality to be encoded into the hashcodes. In this paper we study the effect of learning both the hyperplanes and the thresholds as part of the same model. Most previous research either learn the hyperplanes assuming a fixed set of thresholds, or vice-versa. In our experiments over two standard image datasets we find statistically significant increases in retrieval effectiveness versus a host of state-of-the-art data-dependent and independent hashing models.

In this paper we show how word embeddings can be used to increase the effectiveness of a state-of-the art Locality Sensitive Hashing (LSH) based first story detection (FSD) system over a standard tweet corpus. Vocabulary mismatch, in which related tweets use different words, is a serious hindrance to the effectiveness of a modern FSD system. In this case, a tweet could be flagged as a first story even if a related tweet, which uses different but synonymous words, was already returned as a first story. In this work, we propose a novel approach to mitigate this problem of lexical variation, based on tweet expansion. In particular, we propose to expand tweets with semantically related paraphrases identified via automatically mined word embeddings over a background tweet corpus. Through experimentation on a large data stream comprised of 50 million tweets, we show that FSD effectiveness can be improved by 9.5% over a state-of-the-art FSD system.

In this paper we propose Regularised Cross-Modal Hashing (RCMH) a new cross-modal hashing model that projects annotation and visual feature descriptors into a common Hamming space. RCMH optimises the hashcode similarity of related data-points in the annotation modality using an iterative three-step hashing algorithm: in the first step each training image is assigned a K-bit hashcode based on hyperplanes learnt at the previous iteration; in the second step the binary bits are smoothed by a formulation of graph regularisation so that similar data-points have similar bits; in the third step a set of binary classifiers are trained to predict the regularised bits with maximum margin. Visual descriptors are projected into the annotation Hamming space by a set of binary classifiers learnt using the bits of the corresponding annotations as labels. RCMH is shown to consistently improve retrieval effectiveness over state-of-the-art baselines.

In this paper we introduce a sparse kernel learning framework for the Continuous Relevance Model (CRM). State-of-the-art image annotation models linearly combine evidence from several different feature types to improve image annotation accuracy. While previous authors have focused on learning the linear combination weights for these features, there has been no work examining the optimal combination of kernels. We address this gap by formulating a sparse kernel learning framework for the CRM, dubbed the SKL-CRM, that greedily selects an optimal combination of kernels. Our kernel learning framework rapidly converges to an annotation accuracy that substantially outperforms a host of state-of-the-art annotation models. We make two surprising conclusions: firstly, if the kernels are chosen correctly, only a very small number of features are required so to achieve superior performance over models that utilise a full suite of feature types; and secondly, the standard default selection of kernels commonly used in the literature is sub-optimal, and it is much better to adapt the kernel choice based on the feature type and image dataset.

We introduce a scheme for optimally allocating multiple bits per hyperplane for Locality Sensitive Hashing (LSH). Existing approaches binarise LSH projections by thresholding at zero yielding a single bit per dimension. We demonstrate that this is a sub-optimal bit allocation approach that can easily destroy the neighbourhood structure in the original feature space. Our proposed method, dubbed Neighbourhood Preserving Quantization (NPQ), assigns multiple bits per hyperplane based upon adaptively learned thresholds. NPQ exploits a pairwise affinity matrix to discretise each dimension such that nearest neighbours in the original feature space fall within the same quantisation thresholds and are therefore assigned identical bits. NPQ is not only applicable to LSH, but can also be applied to any low-dimensional projection scheme. Despite using half the number of hyperplanes, NPQ is shown to improve LSH-based retrieval accuracy by up to 65% compared to the state-of-the-art.

In this paper we introduce a new form of the Continuous Relevance Model (the BSCRM) that captures the correlation between tags in a formal and consistent manner. We apply a beam search algorithm to find a near optimal set of mutually correlated tags for an image in a time that is linear in the depth of the search tree. We conduct an examination of the model performance under different kernels for the representation of the image feature distributions and suggest a method of adapting the kernel to the dataset. BS-CRM with a Minkowski kernel is found to significantly increase recall by 42% and precision by 38% over the original CRM model and outperforms more recent baselines on the standard Corel 5k dataset.

This paper reports on work which fuses three different appearance models to enable robust tracking of multiple objects on the basis of colour. Short-term variation in object colour is modelled non-parametrically using adaptive binning histograms. Appearance changes at intermediate time scales are represented by semi-parametric (Gaussian mixture) models while a parametric subspace method (Robust PCA) is employed to model long term stable appearance. Fusion of the three models is achieved through particle filtering and the Democratic integration method. It is shown how robust estimation and adaptation of the models both individually and in combination results in improved visual tracking accuracy. 1 Introduction and Related Work Appearance models play a vital role in visual tracking of objects. They must remain robust to confounding factors such as noise, occlusions, lighting changes, and background variation while adapting to appearance changes caused by motions and deformations of