30 November 2020 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Deep-learning object detection methods that are designed for computer vision applications tend to under-perform when applied to remote sensing data. This is because, contrary to computer vision, in remote sensing training data are harder to collect and targets can be very small, occupying only a few pixels in the entire image, and exhibit arbitrary perspective transformations. Detection performance can improve by fusing data from multiple remote sensing modalities, including RGB, IR, hyper-spectral, multi-spectral, synthetic aperture radar, and LiDAR, to name a few. In this work, we propose YOLOrs: a new convolutional neural network, specifically designed for realtime object detection in multimodal remote sensing imagery. YOLOrs can detect objects at multiple scales, with smaller receptive fields to account for small targets, as well as predict target orientations. In addition, YOLOrs introduces a novel midlevel fusion architecture that renders it applicable to multimodal aerial imagery. Our experimental studies compare YOLOrs with contemporary alternatives and corroborate its merits.
10 July 2014 in IEEE Transactions on Signal Processing
We describe ways to define and calculate L 1 -norm signal subspaces that are less sensitive to outlying data than L 2 -calculated subspaces. We start with the computation of the L 1 maximum-projection principal component of a data matrix containing N signal samples of dimension D. We show that while the general problem is formally NP-hard in asymptotically large N, D, the case of engineering interest of fixed dimension D and asymptotically large sample size N is not. In particular, for the case where the sample size is less than the fixed dimension , we present in explicit form an optimal algorithm of computational cost 2 N . For the case N ≥ D, we present an optimal algorithm of complexity O(N D ). We generalize to multiple L 1 -max-projection components and present an explicit optimal L 1 subspace calculation algorithm of complexity O(N DK-K+1 ) where K is the desired number of L 1 principal components (subspace rank). We conclude with illustrations of L 1 -subspace signal processing in the fields of data dimensionality reduction, direction-of-arrival estimation, and image conditioning/restoration.
30 March 2018 in IEEE Transactions on Signal Processing
L1-norm Principal-Component Analysis (L1-PCA) of real-valued data has attracted significant research interest over the past decade. L1-PCA of complex-valued data remains to date unexplored despite the many possible applications (in communication systems, for example). In this paper, we establish theoretical and algorithmic foundations of L1-PCA of complex-valued data matrices. Specifically, we first show that, in contrast to the real-valued case for which an optimal polynomial-cost algorithm was recently reported by Markopoulos, Karystinos, and Pados, complex L1-PCA is formally NP-hard. Then, casting complex L1-PCA as a unimodular optimization problem, we present the first two suboptimal algorithms in the literature for its solution. Extensive experimental studies included in this paper illustrate the sturdy resistance of complex L1-PCA against faulty measurements/outliers in the processed data.
22 November 2019 in IEEE Access
Tucker decomposition is a standard multi-way generalization of Principal-Component Analysis (PCA), appropriate for processing tensor data. Similar to PCA, Tucker decomposition has been shown to be sensitive against faulty data, due to its L2-norm-based formulation which places squared emphasis to peripheral/outlying entries. In this work, we explore L1-Tucker, an L1-norm based reformulation of Tucker decomposition, and present two algorithms for its solution, namely L1-norm Higher-Order Singular Value Decomposition (L1-HOSVD) and L1-norm Higher-Order Orthogonal Iterations (L1-HOOI). The proposed algorithms are accompanied by complexity and convergence analysis. Our numerical studies on tensor reconstruction and classification corroborate that L1-Tucker decomposition, implemented by means of the proposed algorithms, attains similar performance to standard Tucker when the processed data are corruption-free, while it exhibits sturdy resistance against heavily corrupted entries.
4 October 2018 in IEEE Journal of Selected Topics in Signal Processing
L1-norm Principal-Component Analysis (L1-PCA) is known to attain sturdy resistance against faulty points (outliers) among the processed data. However, computing the L1-PCA of large datasets, with high number of measurements and/or dimensions, may be computationally impractical; in such cases, incremental solutions could be preferred. At the same time, in many applications it is desired to track the signal subspace via principal component adaptation. In this work, we present new methods for both incremental and adaptive L1-PCA. Our first algorithm computes L1-PCA incrementally, processing one measurement at a time and performing online rejection of possible outliers; due to its low computational and storage cost, this algorithm is appropriate for application to both large and streaming datasets. Our second algorithm combines the merits of the first one with the additional ability to track changes in the nominal signal subspace. The proposed algorithms are evaluated with experimental studies on subspace estimation/tracking, video surveillance, image conditioning, and direction-of-arrival (DoA) estimation/tracking.
8 January 2018 in IEEE Signal Processing Letters
This letter studies the rank-1 L1-norm-based TUCKER2 (L1-TUCKER2) decomposition of 3-way tensors. First, we prove that the problem is formally NP-hard. Then, we derive the first two algorithms in the literature for its exact solution. Our algorithms are accompanied by formal complexity analysis. Finally, we conduct numerical studies to compare the performance of exact L1-TUCKER2 (proposed) with standard HOSVD, HOOI, GLRAM, PCA, L1-PCA, and TPCA-L1. In our numerical studies, L1-TUCKER2 outperforms (in tensor approximation) all the aforementioned counterparts when the processed data are outlier corrupted.