We explore the effect of dimensionality on the nearest neighbor problem. We show that under a broad set of conditions (much broader than independent and identically distributed dimensions), as dimensionality increases, the distance to the nearest data point approaches the distance to the farthest data point. To provide a practical perspective, we present empirical results on both real and synthetic data sets that demonstrate that this effect can occur for as few as 10-15 dimensions. These results should not be interpreted to mean that high-dimensional indexing is never meaningful; we illustrate this point by identifying some high-dimensional workloads for which this effect does not occur. However, our results do emphasize that the methodology used almost universally in the database literature to evaluate high-dimensional indexing techniques is flawed, and should be modified. In particular, most such techniques proposed in the literature are not evaluated versus simple linear scan, and are evaluated over workloads for which nearest neighbor is not meaningful. Often, even the reported experiments, when analyzed carefully, show that linear scan would outperform the techniques being proposed on the workloads studied in high (10-15) dimensionality!.

When is nearest neighbor meaningful

We present a new algorithm for manifold learning and nonlinear dimensionality reduction. Based on a set of unorganized data points sampled with noise from a parameterized manifold, the local geometry of the manifold is learned by constructing an approximation for the tangent space at each data point, and those tangent spaces are then aligned to give the global coordinates of the data points with respect to the underlying manifold. We also present an error analysis of our algorithm showing that reconstruction errors can be quite small in some cases. We illustrate our algorithm using curves and surfaces both in two-dimensional/three-dimensional (2D/3D) Euclidean spaces and in higher-dimensional Euclidean spaces. We also address several theoretical and algorithmic issues for further research and improvements.

/pdf/principal-manifolds-and-nonlinear-dimensionality-reduction-442mrhilga.pdf

Principal Manifolds and Nonlinear Dimensionality Reduction via Tangent Space Alignment

The major change in the second edition of this book is the addition of a new chapter on probabilistic retrieval. This chapter has been included because I think this is one of the most interesting and active areas of research in information retrieval. There are still many problems to be solved so I hope that this particular chapter will be of some help to those who want to advance the state of knowledge in this area. All the other chapters have been updated by including some of the more recent work on the topics covered. In preparing this new edition I have benefited from discussions with Bruce Croft, The material of this book is aimed at advanced undergraduate information (or computer) science students, postgraduate library science students, and research workers in the field of IR. Some of the chapters, particularly Chapter 6 * , make simple use of a little advanced mathematics. However, the necessary mathematical tools can be easily mastered from numerous mathematical texts that now exist and, in any case, references have been given where the mathematics occur. I had to face the problem of balancing clarity of exposition with density of references. I was tempted to give large numbers of references but was afraid they would have destroyed the continuity of the text. I have tried to steer a middle course and not compete with the Annual Review of Information Science and Technology. Normally one is encouraged to cite only works that have been published in some readily accessible form, such as a book or periodical. Unfortunately, much of the interesting work in IR is contained in technical reports and Ph.D. theses. For example, most the work done on the SMART system at Cornell is available only in reports. Luckily many of these are now available through the National Technical Information Service (U.S.) and University Microfilms (U.K.). I have not avoided using these sources although if the same material is accessible more readily in some other form I have given it preference. I should like to acknowledge my considerable debt to many people and institutions that have helped me. Let me say first that they are responsible for many of the ideas in this book but that only I wish to be held responsible. My greatest debt is to Karen Sparck Jones who taught me to research information retrieval as an experimental science. Nick Jardine and Robin …

Information retrieval

Speech emotion recognition using deep 1D & 2D CNN LSTM networks

We address the problem of tracking and recognizing faces in real-world, noisy videos. We track faces using a tracker that adaptively builds a target model reflecting changes in appearance, typical of a video setting. However, adaptive appearance trackers often suffer from drift, a gradual adaptation of the tracker to non-targets. To alleviate this problem, our tracker introduces visual constraints using a combination of generative and discriminative models in a particle filtering framework. The generative term conforms the particles to the space of generic face poses while the discriminative one ensures rejection of poorly aligned targets. This leads to a tracker that significantly improves robustness against abrupt appearance changes and occlusions, critical for the subsequent recognition phase. Identity of the tracked subject is established by fusing pose-discriminant and person-discriminant features over the duration of a video sequence. This leads to a robust video-based face recognizer with state-of-the-art recognition performance. We test the quality of tracking and face recognition on real-world noisy videos from YouTube as well as the standard Honda/UCSD database. Our approach produces successful face tracking results on over 80% of all videos without video or person-specific parameter tuning. The good tracking performance induces similarly high recognition rates: 100% on Honda/UCSD and over 70% on the YouTube set containing 35 celebrities in 1500 sequences.

/pdf/face-tracking-and-recognition-with-visual-constraints-in-253utyq21i.pdf

Face tracking and recognition with visual constraints in real-world videos

Effective training of convolutional neural networks for face-based gender and age prediction

This paper addresses the problem of image features selection for pedestrian gender recognition. Hand-crafted features (such as HOG) are compared with learned features which are obtained by training convolutional neural networks. The comparison is performed on the recently created collection of versatile pedestrian datasets which allows us to evaluate the impact of dataset properties on the performance of features. The study shows that hand-crafted and learned features perform equally well on small-sized homogeneous datasets. However, learned features significantly outperform hand-crafted ones in the case of heterogeneous and unfamiliar (unseen) datasets. Our best model which is based on learned features obtains 79% average recognition rate on completely unseen datasets. We also show that a relatively small convolutional neural network is able to produce competitive features even with little training data.

/pdf/learned-vs-hand-crafted-features-for-pedestrian-gender-363676ycb9.pdf

Learned vs. Hand-Crafted Features for Pedestrian Gender Recognition

https://www.eurecom.fr/fr/publication/4768/download/mm-publi-4768.pdf

Minimalistic CNN-based ensemble model for gender prediction from face images

This paper proposes a novel content-based image retrieval scheme for image copy identification. Its goal is to detect matches between a set of doubtful images and the ones stored in the database of the legal holders of the photographies. If an image was stolen and used to create a pirated copy, it tries to identify from which original image that copy was created. The image recognition scheme is based on local differential descriptors. Therefore, the matching process takes into account a large set of variations that might have been applied to stolen images in order to create pirated copies. The high cost and the complexity of this image recognition scheme requires a very efficient retrieval process since many individual queries must be executed before being able to construct the final result. This paper therefore proposes to use a novel search method that trades the precision of each individual search for reduced query execution time. This imprecision has only little impact on the overall recognition performance since the final result is a consolidation of many partial results. However, it dramatically accelerates queries. This result has then been corroborated by a theoretically study. Experiments show the efficiency and the robustness of the proposed scheme.

Robust content-based image searches for copyright protection

This work describes our solution in the second edition of the ChaLearn LAP competition on Apparent Age Estimation. Starting from a pretrained version of the VGG-16 convolutional neural network for face recognition, we train it on the huge IMDB-Wiki dataset for biological age estimation and then fine-tune it for apparent age estimation using the relatively small competition dataset. We show that the precise age estimation of children is the cornerstone of the competition. Therefore, we integrate a separate "children" VGG-16 network for apparent age estimation of children between 0 and 12 years old in our final solution. The "children" network is fine-tuned from the "general" one. We employ different age encoding strategies for training "general" and "children" networks: the soft one (label distribution encoding) for the "general" network and the strict one (0/1 classification encoding) for the "children" network. Finally, we highlight the importance of the state-of-the-art face detection and face alignment for the final apparent age estimation. Our resulting solution wins the 1st place in the competition significantly outperforming the runner-up.

/pdf/apparent-age-estimation-from-face-images-combining-general-4ta291exoq.pdf

Sid-Ahmed Berrani

Papers

Effective training of convolutional neural networks for face-based gender and age prediction

Learned vs. Hand-Crafted Features for Pedestrian Gender Recognition

Minimalistic CNN-based ensemble model for gender prediction from face images

Robust content-based image searches for copyright protection

Apparent Age Estimation from Face Images Combining General and Children-Specialized Deep Learning Models