Kobus Barnard

Matching words and pictures

We present a new approach for modeling multi-modal data sets, focusing on the specific case of segmented images with associated text. Learning the joint distribution of image regions and words has many applications. We consider in detail predicting words associated with whole images (auto-annotation) and corresponding to particular image regions (region naming). Auto-annotation might help organize and access large collections of images. Region naming is a model of object recognition as a process of translating image regions to words, much as one might translate from one language to another. Learning the relationships between image regions and semantic correlates (words) is an interesting example of multi-modal data mining, particularly because it is typically hard to apply data mining techniques to collections of images. We develop a number of models for the joint distribution of image regions and words, including several which explicitly learn the correspondence between regions and words. We study multi-modal and correspondence extensions to Hofmanns hierarchical clustering/aspect model, a translation model adapted from statistical machine translation (Brown et al.), and a multi-modal extension to mixture of latent Dirichlet allocation (MoM-LDA). All models are assessed using a large collection of annotated images of real scenes. We study in depth the difficult problem of measuring performance. For the annotation task, we look at prediction performance on held out data. We present three alternative measures, oriented toward different types of task. Measuring the performance of correspondence methods is harder, because one must determine whether a word has been placed on the right region of an image. We can use annotation performance as a proxy measure, but accurate measurement requires hand labeled data, and thus must occur on a smaller scale. We show results using both an annotation proxy, and manually labeled data.

Object Recognition as Machine Translation: Learning a Lexicon for a Fixed Image Vocabulary

We describe a model of object recognition as machine translation. In this model, recognition is a process of annotating image regions with words. Firstly, images are segmented into regions, which are classified into region types using a variety of features. A mapping between region types and keywords supplied with the images, is then learned, using a method based around EM. This process is analogous with learning a lexicon from an aligned bitext. For the implementation we describe, these words are nouns taken from a large vocabulary. On a large test set, the method can predict numerous words with high accuracy. Simple methods identify words that cannot be predicted well. We show how to cluster words that individually are difficult to predict into clusters that can be predicted well -- for example, we cannot predict the distinction between train and locomotive using the current set of features, but we can predict the underlying concept. The method is trained on a substantial collection of images. Extensive experimental results illustrate the strengths and weaknesses of the approach.

Learning the semantics of words and pictures

We present a statistical model for organizing image collections which integrates semantic information provided by associate text and visual information provided by image features. The model is very promising for information retrieval tasks such as database browsing and searching for images based on text and/or image features. Furthermore, since the model learns relationships between text and image features, it can be used for novel applications such as associating words with pictures, and unsupervised learning for object recognition.

A comparison of computational color constancy algorithms. I: Methodology and experiments with synthesized data

We introduce a context for testing computational color constancy, specify our approach to the implementation of a number of the leading algorithms, and report the results of three experiments using synthesized data. Experiments using synthesized data are important because the ground truth is known, possible confounds due to camera characterization and pre-processing are absent, and various factors affecting color constancy can be efficiently investigated because they can be manipulated individually and precisely. The algorithms chosen for close study include two gray world methods, a limiting case of a version of the Retinex method, a number of variants of Forsyths gamut-mapping method, Cardei et al.s neural net method, and Finlayson et al.s color by correlation method. We investigate the ability of these algorithms to make estimates of three different color constancy quantities: the chromaticity of the scene illuminant, the overall magnitude of that illuminant, and a corrected, illumination invariant, image. We consider algorithm performance as a function of the number of surfaces in scenes generated from reflectance spectra, the relative effect on the algorithms of added specularities, and the effect of subsequent clipping of the data. All data is available on-line at http://www.cs.sfu.ca/(tilde)color/data, and implementations for most of the algorithms are also available (http://www.cs.sfu.ca/(tilde)color/code).

A comparison of computational color constancy Algorithms. II. Experiments with image data

We test a number of the leading computational color constancy algorithms using a comprehensive set of images. These were of 33 different scenes under 11 different sources representative of common illumination conditions. The algorithms studied include two gray world methods, a version of the Retinex method, several variants of Forsyths gamut-mapping method, Cardei et al.s neural net method, and Finlayson et al.s Color by Correlation method. We discuss a number of issues in applying color constancy ideas to image data, and study in depth the effect of different preprocessing strategies. We compare the performance of the algorithms on image data with their performance on synthesized data. All data used for this study are available online at http://www.cs.sfu.ca/(tilde)color/data, and implementations for most of the algorithms are also available (http://www.cs.sfu.ca/(tilde)color/code). Experiments with synthesized data (part one of this paper) suggested that the methods which emphasize the use of the input data statistics, specifically color by correlation and the neural net algorithm, are potentially the most effective at estimating the chromaticity of the scene illuminant. Unfortunately, we were unable to realize comparable performance on real images. Here exploiting pixel intensity proved to be more beneficial than exploiting the details of image chromaticity statistics, and the three-dimensional (3-D) gamut-mapping algorithms gave the best performance.

A statistical model for general contextual object recognition

We consider object recognition as the process of attaching meaningful labels to specific regions of an image, and propose a model that learns spatial relationships between objects. Given a set of images and their associated text (e.g. keywords, captions, descriptions), the objective is to segment an image, in either a crude or sophisticated fashion, then to find the proper associations between words and regions. Previous models are limited by the scope of the representation. In particular, they fail to exploit spatial context in the images and words. We develop a more expressive model that takes this into account. We formulate a spatially consistent probabilistic mapping between continuous image feature vectors and the supplied word tokens. By learning both word-to-region associations and object relations, the proposed model augments scene segmentations due to smoothing implicit in spatial consistency. Context introduces cycles to the undirected graph, so we cannot rely on a straightforward implementation of the EM algorithm for estimating the model parameters and densities of the unknown alignment variables. Instead, we develop an approximate EM algorithm that uses loopy belief propagation in the inference step and iterative scaling on the pseudo-likelihood approximation in the parameter update step. The experiments indicate that our approximate inference and learning algorithm converges to good local solutions. Experiments on a diverse array of images show that spatial context considerably improves the accuracy of object recognition. Most significantly, spatial context combined with a nonlinear discrete object representation allows our models to cope well with over-segmented scenes.

Is Machine Colour Constancy Good Enough

This paper presents a negative result: current machine colour constancy algorithms are not good enough for colour-based object recognition. This result has surprised us since we have previously used the better of these algorithms successfully to correct the colour balance of images for display. Colour balancing has been the typical application of colour constancy, rarely has it been actually put to use in a computer vision system, so our goal was to show how well the various methods would do on an obvious machine colour vision task, namely, object recognition. Although all the colour constancy methods we tested proved insufficient for the task, we consider this an important finding in itself. In addition we present results showing the correlation between colour constancy performance and object recognition performance, and as one might expect, the better the colour constancy the better the recognition rate.

Color Constancy for Scenes with Varying Illumination

We present an algorithm which uses information from both surface reflectance and illumination variation to solve for color constancy. Most color constancy algorithms assume that the illumination across a scene is constant, but this is very often not valid for real images. The method presented in this work identifies and removes the illumination variation, and in addition uses the variation to constrain the solution. The constraint is applied conjunctively to constraints found from surface reflectances. Thus the algorithm can provide good color constancy when there is sufficient variation in surface reflectances, or sufficient illumination variation, or a combination of both. We present the results of running the algorithm on several real scenes, and the results are very encouraging.

Estimating the scene illumination chromaticity by using a neural network

A neural network can learn color constancy, defined here as the ability to estimate the chromaticity of a scenes overall illumination. We describe a multilayer neural network that is able to recover the illumination chromaticity given only an image of the scene. The network is previously trained by being presented with a set of images of scenes and the chromaticities of the corresponding scene illuminants. Experiments with real images show that the network performs better than previous color constancy methods. In particular, the performance is better for images with a relatively small number of distinct colors. The method has application to machine vision problems such as object recognition, where illumination-independent color descriptors are required, and in digital photography, where uncontrolled scene illumination can create an unwanted color cast in a photograph.

Color constancy under varying illumination

Illumination is rarely constant in intensity or color throughout a scene. Multiple light sources with different spectra-sun and sky, direct and interreflected light-are the norm. Nonetheless, almost all color constancy algorithms assume that the spectrum of the incident illumination remains constant across the scene. We assume the converse, that illumination does vary, in developing a new algorithm for color constancy. Rather than creating difficulties, varying illumination is in fact a very powerful constraint. Indeed tests of our algorithm using real images of an office scene show excellent results.<<ETX>>