arXiv subjects

Aggelos K. Katsaggelos

Publications and source records attributed to Aggelos K. Katsaggelos.

67 records · Page 4Linked to original sources

arXiv2019

A Single Video Super-Resolution GAN for Multiple Downsampling Operators based on Pseudo-Inverse Image Formation Models

The popularity of high and ultra-high definition displays has led to the need for methods to improve the quality of videos already obtained at much lower resolutions. Current Video Super-Resolution methods are not robust to mismatch between training and testing degradation models since they are trained against a single degradation model (usually bicubic downsampling). This causes their performance to deteriorate in real-life applications. At the same time, the use of only the Mean Squared Error during learning causes the resulting images to be too smooth. In this work we propose a new Convolutional Neural Network for video super resolution which is robust to multiple degradation models. During training, which is performed on a large dataset of scenes with slow and fast motions, it uses the pseudo-inverse image formation model as part of the network architecture in conjunction with perceptual losses, in addition to a smoothness constraint that eliminates the artifacts originating from these perceptual losses. The experimental validation shows that our approach outperforms current state-of-the-art methods and is robust to multiple degradations.

cs.CV↗

arXiv2018

Generative Adversarial Networks and Perceptual Losses for Video Super-Resolution

Video super-resolution (VSR) has become one of the most critical problems in video processing. In the deep learning literature, recent works have shown the benefits of using adversarial-based and perceptual losses to improve the performance on various image restoration tasks; however, these have yet to be applied for video super-resolution. In this work, we propose a Generative Adversarial Network(GAN)-based formulation for VSR. We introduce a new generator network optimized for the VSR problem, named VSRResNet, along with a new discriminator architecture to properly guide VSRResNet during the GAN training. We further enhance our VSR GAN formulation with two regularizers, a distance loss in feature-space and pixel-space, to obtain our final VSRResFeatGAN model. We show that pre-training our generator with the Mean-Squared-Error loss only quantitatively surpasses the current state-of-the-art VSR models. Finally, we employ the PercepDist metric (Zhang et al., 2018) to compare state-of-the-art VSR models. We show that this metric more accurately evaluates the perceptual quality of SR solutions obtained from neural networks, compared with the commonly used PSNR/SSIM metrics. Finally, we show that our proposed model, the VSRResFeatGAN model, outperforms current state-of-the-art SR models, both quantitatively and qualitatively.

cs.CV↗

arXiv2018

Neuroimaging Modality Fusion in Alzheimer's Classification Using Convolutional Neural Networks

Automated methods for Alzheimer's disease (AD) classification have the potential for great clinical benefits and may provide insight for combating the disease. Machine learning, and more specifically deep neural networks, have been shown to have great efficacy in this domain. These algorithms often use neurological imaging data such as MRI and PET, but a comprehensive and balanced comparison of these modalities has not been performed. In order to accurately determine the relative strength of each imaging variant, this work performs a comparison study in the context of Alzheimer's dementia classification using the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. Furthermore, this work analyzes the benefits of using both modalities in a fusion setting and discusses how these data types may be leveraged in future AD studies using deep learning.

cs.LG↗

arXiv2018

Computational multifocal microscopy

Despite recent advances, high performance single-shot 3D microscopy remains an elusive task. By introducing designed diffractive optical elements (DOEs), one is capable of converting a microscope into a 3D "kaleidoscope", in which case the snapshot image consists of an array of tiles and each tile focuses on different depths. However, the acquired multifocal microscopic (MFM) image suffers from multiple sources of degradation, which prevents MFM from further applications. We propose a unifying computational framework which simplifies the imaging system and achieves 3D reconstruction via computation. Our optical configuration omits chromatic correction grating and redesigns the multifocal grating to enlarge the tracking area. Our proposed setup features only one single grating in addition to a regular microscope. The aberration correction, along with Poisson and background denoising, are incorporated in our deconvolution-based fully-automated algorithm, which requires no empirical parameter-tuning. In experiments, we achieve the spatial resolutions of $0.35$um (lateral) and $0.5$um (axial), which are comparable to the resolution that can be achieved with confocal deconvolution microscopy. We demonstrate a 3D video of moving bacteria recorded at $25$ frames per second using our proposed computational multifocal microscopy technique.

physics.optics↗

arXiv2018

3D Snapshot Microscopy of Extended Objects

Volumetric biological imaging often involves compromising high temporal resolution at the expense of high spatial resolution when popular scanning methods are used to capture 3D information. We introduce an integrated experimental and image reconstruction method for capturing dynamic 3D fluorescent extended objects as a series of synchronously measured 3D snapshots taken at the frame rate of the imaging camera. We employ multifocal microscopy (MFM) to simultaneously image at 25 focal planes and process this depth-encoded image to recover the 3D structure of extended objects, such as bacteria, using a sparsity-based reconstruction approach. The combined experimental and computational method produces image quality similar to confocal microscopy in a fraction of the acquisition time. In addition, our computational image reconstruction approach allows a simplified MFM optical design by correcting aberrations using the measured response to point sources. This "compressive" MFM acquisition and reconstruction method, where an image volume with roughly 8 million voxels is recovered from a single 1-megapixel captured image, enables straightforward study of dynamic processes in 3D, and as a simultaneous snapshot advances the state of the art in dynamic 3D microscopy.

physics.ins-det↗

arXiv2018

DIRECT: Deep Discriminative Embedding for Clustering of LIGO Data

In this paper, benefiting from the strong ability of deep neural network in estimating non-linear functions, we propose a discriminative embedding function to be used as a feature extractor for clustering tasks. The trained embedding function transfers knowledge from the domain of a labeled set of morphologically-distinct images, known as classes, to a new domain within which new classes can potentially be isolated and identified. Our target application in this paper is the Gravity Spy Project, which is an effort to characterize transient, non-Gaussian noise present in data from the Advanced Laser Interferometer Gravitational-wave Observatory, or LIGO. Accumulating large, labeled sets of noise features and identifying of new classes of noise lead to a better understanding of their origin, which makes their removal from the data and/or detectors possible.

cs.LG↗

arXiv2018

Efficient Video Object Segmentation via Network Modulation

Video object segmentation targets at segmenting a specific object throughout a video sequence, given only an annotated first frame. Recent deep learning based approaches find it effective by fine-tuning a general-purpose segmentation model on the annotated frame using hundreds of iterations of gradient descent. Despite the high accuracy these methods achieve, the fine-tuning process is inefficient and fail to meet the requirements of real world applications. We propose a novel approach that uses a single forward pass to adapt the segmentation model to the appearance of a specific object. Specifically, a second meta neural network named modulator is learned to manipulate the intermediate layers of the segmentation network given limited visual and spatial information of the target object. The experiments show that our approach is 70times faster than fine-tuning approaches while achieving similar accuracy.

cs.CV↗

arXiv2017

Deep Fully-Connected Networks for Video Compressive Sensing

In this work we present a deep learning framework for video compressive sensing. The proposed formulation enables recovery of video frames in a few seconds at significantly improved reconstruction quality compared to previous approaches. Our investigation starts by learning a linear mapping between video sequences and corresponding measured frames which turns out to provide promising results. We then extend the linear formulation to deep fully-connected networks and explore the performance gains using deeper architectures. Our analysis is always driven by the applicability of the proposed framework on existing compressive video architectures. Extensive simulations on several video sequences document the superiority of our approach both quantitatively and qualitatively. Finally, our analysis offers insights into understanding how dataset sizes and number of layers affect reconstruction performance while raising a few points for future investigation. Code is available at Github: https://github.com/miliadis/DeepVideoCS

cs.CV↗

arXiv2017

Photometric Stereo by UV-Induced Fluorescence to Detect Protrusions on Georgia O'Keeffe's Paintings

A significant number of oil paintings produced by Georgia O'Keeffe (1887-1986) show surface protrusions of varying width, up to several hundreds of microns. These protrusions are similar to those described in the art conservation literature as metallic soaps. Since the presence of these protrusions raises questions about the state of conservation and long-term prospects for deterioration of these artworks, a 3D-imaging technique, photometric stereo using ultraviolet illumination, was developed for the long-term monitoring of the surface-shape of the protrusions and the surrounding paint. Because the UV fluorescence response of painting materials is isotropic, errors typically caused by non-Lambertian (anisotropic) specularities when using visible reflected light can be avoided providing a more accurate estimation of shape. As an added benefit, fluorescence provides additional contrast information contributing to materials characterization. The developed methodology aims to detect, characterize, and quantify the distribution of micro-protrusions and their development over the surface of entire artworks. Combined with a set of analytical in-situ techniques, and computational tools, this approach constitutes a novel methodology to investigate the selective distribution of protrusions in correlation with the composition of painting materials at the macro-scale. While focused on O'Keeffe's paintings as a case study, we expect the proposed approach to have broader significance by providing a non-invasive protocol to the conservation community to probe topological changes for any relatively flat painted surface of an artwork, and more specifically to monitor the dynamic formation of protrusions, in relation to paint composition and modifications of environmental conditions, loans, exhibitions and storage over the long-term.

cs.GR↗

arXiv2017

Robust and Low-Rank Representation for Fast Face Identification with Occlusions

In this paper we propose an iterative method to address the face identification problem with block occlusions. Our approach utilizes a robust representation based on two characteristics in order to model contiguous errors (e.g., block occlusion) effectively. The first fits to the errors a distribution described by a tailored loss function. The second describes the error image as having a specific structure (resulting in low-rank in comparison to image size). We will show that this joint characterization is effective for describing errors with spatial continuity. Our approach is computationally efficient due to the utilization of the Alternating Direction Method of Multipliers (ADMM). A special case of our fast iterative algorithm leads to the robust representation method which is normally used to handle non-contiguous errors (e.g., pixel corruption). Extensive results on representative face databases (in constrained and unconstrained environments) document the effectiveness of our method over existing robust representation methods with respect to both identification rates and computational time. Code is available at Github, where you can find implementations of the F-LR-IRNNLS and F-IRNNLS (fast version of the RRC) : https://github.com/miliadis/FIRC

cs.CV↗

arXiv2016

Compressive Holographic Video

Compressed sensing has been discussed separately in spatial and temporal domains. Compressive holography has been introduced as a method that allows 3D tomographic reconstruction at different depths from a single 2D image. Coded exposure is a temporal compressed sensing method for high speed video acquisition. In this work, we combine compressive holography and coded exposure techniques and extend the discussion to 4D reconstruction in space and time from one coded captured image. In our prototype, digital in-line holography was used for imaging macroscopic, fast moving objects. The pixel-wise temporal modulation was implemented by a digital micromirror device. In this paper we demonstrate $10\times$ temporal super resolution with multiple depths recovery from a single image. Two examples are presented for the purpose of recording subtle vibrations and tracking small particles within 5 ms.

cs.CV↗

arXiv2016

DeepBinaryMask: Learning a Binary Mask for Video Compressive Sensing

In this paper, we propose a novel encoder-decoder neural network model referred to as DeepBinaryMask for video compressive sensing. In video compressive sensing one frame is acquired using a set of coded masks (sensing matrix) from which a number of video frames is reconstructed, equal to the number of coded masks. The proposed framework is an end-to-end model where the sensing matrix is trained along with the video reconstruction. The encoder learns the binary elements of the sensing matrix and the decoder is trained to recover the unknown video sequence. The reconstruction performance is found to improve when using the trained sensing mask from the network as compared to other mask designs such as random, across a wide variety of compressive sensing reconstruction algorithms. Finally, our analysis and discussion offers insights into understanding the characteristics of the trained mask designs that lead to the improved reconstruction quality.

cs.CV↗

arXiv2011

Sparse Bayesian Methods for Low-Rank Matrix Estimation

Recovery of low-rank matrices has recently seen significant activity in many areas of science and engineering, motivated by recent theoretical results for exact reconstruction guarantees and interesting practical applications. A number of methods have been developed for this recovery problem. However, a principled method for choosing the unknown target rank is generally not provided. In this paper, we present novel recovery algorithms for estimating low-rank matrices in matrix completion and robust principal component analysis based on sparse Bayesian learning (SBL) principles. Starting from a matrix factorization formulation and enforcing the low-rank constraint in the estimates as a sparsity constraint, we develop an approach that is very effective in determining the correct rank while providing high recovery performance. We provide connections with existing methods in other similar problems and empirical results and comparisons with current state-of-the-art methods that illustrate the effectiveness of this approach.

stat.ML↗