Reconstruction of Multi-Temporal Satellite Imagery by Coupling Variational Segmentation and Radiometric Analysis

Digital images, and in particular satellite images acquired by different sensors, may present defects due to many causes. Since 2013, the Landsat 7 mission has been affected by a well-known issue related to the malfunctioning of the Scan Line Corrector producing very characteristic strips of missing data in the imagery bands. Within the vast and interdisciplinary image reconstruction application field, many works have been presented in the last few decades to tackle the specific Landsat 7 gap-filling problem. This work proposes another contribution in this field presenting an original procedure based on a variational image segmentation model coupled with radiometric analysis to reconstruct damaged images acquired in a multi-temporal scenario, typical in satellite remote sensing. The key idea is to exploit some specific features of the Mumford–Shah variational model for image segmentation in order to ease the detection of homogeneous regions which will then be used to form a set of coherent data necessary for the radiometric reconstruction of damaged regions. Two reconstruction approaches are presented and applied to SLC-off Landsat 7 data. One approach is based on the well-known histogram matching transformation, the other approach is based on eigendecomposition of the bands covariance matrix and on the sampling from Gaussian distributions. The performance of the procedure is assessed by application to artificially damaged images for self-validation testing. Both of the proposed reconstruction approaches had led to remarkable results. An application to very high resolution WorldView-3 data shows how the procedure based on variational segmentation allows an effective reconstruction of images presenting a great level of geometric complexity.

Training Set Enlargement Using Binary Weighted Interpolation Maps for the Single Sample per Person Problem in Face Recognition

We propose a method of enlarging the training dataset for a single-sample-per-person (SSPP) face recognition problem. The appearance of the human face varies greatly, owing to various intrinsic and extrinsic factors. In order to build a face recognition system that can operate robustly in an uncontrolled, real environment, it is necessary for the algorithm to learn various images of the same person. However, owing to limitations in the collection of facial image data, only one sample can typically be obtained, causing difficulties in the performance and usability of the method. This paper proposes a method that analyzes the changes in pixels in face images associated with variations by extracting the binary weighted interpolation map (B-WIM) from neutral and variational images in the auxiliary set. Then, a new variational image for the query image is created by combining the given query (neutral) image and the variational image of the auxiliary set based on the B-WIM. As a result of performing facial recognition comparison experiments on SSPP training data for various facial-image databases, the proposed method shows superior performance compared with other methods.

A Multi-Scale Compressed Sensing Algorithm Based on Variational Mode

The compressed sensing algorithm based on the hybrid sparse base (TFWBST+wave atom) usually uses two kinds of image sparse transformations to realize the sparse representation of structure and texture respectively. However, due to the lack of constraints on image texture and structure and the lack of orthogonality of the two sparse bases, the sparse coefficient of structure and the sparse coefficient of texture after transformation are often not good enough to reflect their respective components, that is, the texture coefficient often loses the detail information of texture. To overcome this phenomenon, this paper combines the compressed sensing algorithm based on hybrid base with the layered variational image decomposition method to form the variational multi-scale compressed sensing, which is to establish the CS image reconstruction model with minimal energy functional. The layered variational image decomposition decomposes image into different feature components by minimizing energy functional. The reconstruction of each layer by compressed sensing algorithm is very suitable for texture and detail reconstruction. In this model, TFWBST transform and wave atom are combined as a joint sparse dictionary, and the image decomposition is carried out under the (BV, G, E) variational framework, which is introduced into multi-scale compressed sensing technology to reconstruct the original image. In this new functional, TFWBST transform and wave atom are used to represent structure and texture respectively, and multiscale (BV, G, E) decomposition which can decompose an image into a sequence of image structure, texture and noise is added for restricting image parts. Experiments show that the new model is very robust for noise, and that can keep edges and textures stably than other multi-scale restoration and reconstruction of images.

Variational Hilbert Quantitative Phase Imaging

Utilizing the refractive index as the endogenous contrast agent to noninvasively study transparent cells is a working principle of emerging quantitative phase imaging (QPI). In this contribution, we propose the Variational Hilbert Quantitative Phase Imaging (VHQPI)—end-to-end purely computational add-on module able to improve performance of a QPI-unit without hardware modifications. The VHQPI, deploying unique merger of tailored variational image decomposition and enhanced Hilbert spiral transform, adaptively provides high quality map of sample-induced phase delay, accepting particularly wide range of input single-shot interferograms (from off-axis to quasi on-axis configurations). It especially promotes high space-bandwidth-product QPI configurations alleviating the spectral overlapping problem. The VHQPI is tailored to deal with cumbersome interference patterns related to detailed locally varying biological objects with possibly high dynamic range of phase and relatively low carrier. In post-processing, the slowly varying phase-term associated with the instrumental optical aberrations is eliminated upon variational analysis to further boost the phase-imaging capabilities. The VHQPI is thoroughly studied employing numerical simulations and successfully validated using static and dynamic cells phase-analysis. It compares favorably with other single-shot phase reconstruction techniques based on the Fourier and Hilbert–Huang transforms, both in terms of visual inspection and quantitative evaluation, potentially opening up new possibilities in QPI.