Claus Bauer

Compact hashing with joint optimization of search accuracy and time

Similarity search, namely, finding approximate nearest neighborhoods, is the core of many large scale machine learning or vision applications. Recently, many research results demonstrate that hashing with compact codes can achieve promising performance for large scale similarity search. However, most of the previous hashing methods with compact codes only model and optimize the search accuracy. Search time, which is an important factor for hashing in practice, is usually not addressed explicitly. In this paper, we develop a new scalable hashing algorithm with joint optimization of search accuracy and search time simultaneously. Our method generates compact hash codes for data of general formats with any similarity function. We evaluate our method using diverse data sets up to 1 million samples (e.g., web images). Our comprehensive results show the proposed method significantly outperforms several state-of-the-art hashing approaches.

Content-based Video Signatures based on Projections of Difference Images

We propose a novel video signature extraction method based on projections of difference images between consecutive video frames. The difference images are projected onto random basis vectors to create a low dimensional bitstream representation of the active content (moving regions) between two video frames. A sequence of these signatures serves to identify the underlying video content in a robust manner. Our experimental results show that the proposed video signature is robust to most common signal processing operations on video content such as compression, resolution scaling, brightness scaling.

Robust video fingerprints based on subspace embedding

We present a novel video fingerprinting method based on subspace embedding. The proposed method is particularly robust against frame-rate conversion attacks and geometric attacks among other attacks including compression and spatial scaling. Using a sliding window, we extract fingerprints from a group of subsequent video frames. For the generation of the fingerprints, we first calculate the basis vectors of a coarse representation of this group of frames using a singular value decomposition (SVD). Then, we project the coarse representation of the video frames onto a subset of the basis vectors. Thus, we obtain a subspace representation of the input video frames. Finally, we extract the fingerprint bits by projecting a temporal average of these representations onto pseudorandom basis vectors. Since the subspace is estimated from the input video data itself, any global attack on video such as rotation would result in a corresponding change in estimated basis vectors thereby preserving the subspace representation. We present experimental results on 250 hrs of video to show the robustness and sensitivity of the proposed signature extraction method.

Audio and video signatures for synchronization

We propose a framework based on signatures extracted from audio and video streams for automatically measuring and maintaining synchronization between the two streams. The audio signature is based on projections of a coarse representation of the spectrogram onto random vectors. The video signature is based on projections of a coarse representation of the difference image between two consecutive frames onto random vectors. The time alignment present at the signature generator between the two streams is recorded by combining audio and video signatures into a combined synchronization signature. At the detector after video and audio streams go through different processing operations, we extract the signatures again. The signatures extracted before and after processing from the audio and the video are compared independently using a Hamming distance based correlator to estimate the relative misalignment introduced due to processing in each of the streams. Then, the estimated relative misalignment between the audio and video streams is used to preserve the same alignment between the streams that was present before processing. Our experimental results show that we can achieve > 93.0% accuracy in synchronization.

Audio Signature Extraction Based on Projections of Spectrograms

Content-based signatures are designed to be a robust bit-stream representation of the content so as to enable content identification even though the original content may go through various signal processing operations. In this paper, we propose a novel content-based audio signature extraction method that captures temporal evolution of the audio spectrum. The proposed method, first, divides the input audio into overlapping chunks and computes a spectrogram for each chunk. Then, it projects each of the spectrograms onto random basis vectors to create a signature that is a low-dimensional bit-stream representation of the corresponding spectrogram. Our experimental results show the robustness and sensitivity of the proposed content-based audio signature extraction method for various signal processing operations on audio content.

Joint optimization of scale factors and Huffman code books for MPEG-4 AAC

This paper addresses the optimization problem of minimizing the distortion subject to a rate constraint for an MPEG-4 Advanced Audio Coding (AAC) encoder. We first develop a mathematical model of the AAC encoding process. In previous work, the joint optimization problem is modeled as a Viterbi search for a cheapest path through a trellis. This method involves an iteration over a Lagrangian multiplier. We improve on this method by deriving a very accurate guess for the value of the final Lagrangian multiplier of the iteration as a function of the Perceptual Entropy of the signal and the given rate constraint. This reduces the complexity of the Trellis Search significantly. Whereas previous methods including the Trellis Search did not provide optimal solutions to the problem of minimizing the distortion subject to a rate constraint, we establish two methods that for the first time solve this problem optimally. Our first method is based on the formulation and solution of a Mixed Integer Linear Program, whereas our second method uses a Dynamic Programming solution that does not rely on the iteration over a Lagrangian multiplier. Based on our optimal methods, we evaluate the performance of the heuristic Two Loop Search (TLS), which is used in most commercial AAC implementations to solve the problem under consideration, and the performance of the Trellis Search.

Video fingerprinting based on moment invariants capturing appearance and motion

In this paper, we propose two video fingerprinting methods that are robust to both geometric and non-geometric modifications on content. Both of the proposed methods are based on computation of moment invariants as features from concentric circular regions. The two methods differ in the way they capture appearance and motion information from video. In one method, we capture motion information by computing a difference image between the current video frame and a temporal average video frame computed from a past window of video frames. This method captures appearance by computing moment invariants from concentric circular regions of a video frame. In the second method, we capture appearance and motion by projecting features onto two sets of basis functions and explicitly capture how the moment invariants change over the regions and over time. We present experimental results on both of these video fingerprinting comparing their performance in terms of robustness against attacks and sensitivity to content.

A Statistical Study of Loss-Delay Tradeoff for RED Queues

Aside from the introduction of many new schemes, the use of TCP-based AQM schemes and in specific RED is anticipated to continue in foreseeable future as the de-facto standard of network congestion control. Therefore, conducting extra research work aiming at improving the performance of RED is still a topic of high interest. In this paper, we present an analytical study aiming at the fine tuning of the RED parameters. Utilizing a statistical analysis approach, we formulate an optimization problem aimed at addressing the loss and delay tradeoff of the RED queuing discipline. We provide a two-phase iterative solution to the problem in order to identify the settings of the RED parameters. We discuss the convergence characteristics of our solution and investigate its low complexity characteristics. Through extensive NS2 experiments, we illustrate the advantages of our proposed optimization approach by comparing its results to those of adaptive RED as well as standard RED with recommended parameter settings.

On improving the collision property of robust hashing based on projections

In this paper, we study the collision property of one of the robust hash functions proposed in [1]. This method was originally proposed for robust hash generation from blocks of image data and is based on projection of image block data on pseudo-random matrices. We show that collision performance of this robust hash function is not optimal when used to extract hash bits from a moment invariants feature matrix for video fingerprinting. We identify that the collision performance of this hash extraction method could be improved if the pseudo-random matrices are selected carefully. We propose two methods that use an offline training set to improve the collision property. Both of the methods attempt to select the matrices that minimize cross-correlation among the projected features. The first method uses an iterative procedure to select the matrices that satisfy a cross-correlation threshold. The second method used Singular Value Decomposition (SVD) of the feature covariance matrix and hence the crosscorrelation of the projected values is zero. We show the improved collision performance of both these methods on the same dataset. Also, we interpret the projection matrices obtained through the SVD procedure and show that they capture appearance and motion information from the moment invariants feature matrix.

Optimal Statistical Tuning of the RED Parameters

Achieving minimal loss while satisfying an acceptable delay profile remains to be an open problem under the RED queuing discipline. In this paper, we present a framework targeted at optimal fine tuning of the RED parameters in order to address such problem. For a given traffic pattern and utilizing a statistical analysis of finite-state Markov chains, we formulate an optimization problem aimed at addressing the loss and delay tradeoff of the RED queuing discipline. Our two-step iterative solution to the problem identifies the optimal settings of the RED parameters. We prove the convergence of our solution and investigate its low complexity characteristics. We apply our framework to a number of generic queuing and TCP scenarios in order to capture loss and delay performance of our algorithms versus buffer capacity and service rate. Based on our results, we argue that our model is capable of optimally addressing the loss-delay tradeoff of RED queues accommodating time-varying traffic profiles.