High-fidelity optical diffraction tomography of multiple scattering samples

We propose an iterative reconstruction scheme for optical diffraction tomography that exploits the split-step nonparaxial (SSNP) method as the forward model in a learning tomography scheme. Compared with the beam propagation method (BPM) previously used in learning tomography (LT-BPM), the improved accuracy of SSNP maximizes the information retrieved from measurements, relying less on prior assumptions about the sample. A rigorous evaluation of learning tomography based on SSNP (LT-SSNP) using both synthetic and experimental measurements confirms its superior performance compared with that of the LT-BPM. Benefiting from the accuracy of SSNP, LT-SSNP can clearly resolve structures that are highly distorted in the LT-BPM. A serious limitation for quantifying the reconstruction accuracy for biological samples is that the ground truth is unknown. To overcome this limitation, we describe a novel method that allows us to compare the performances of different reconstruction schemes by using the discrete dipole approximation to generate synthetic measurements. Finally, we explore the capacity of learning approaches to enable data compression by reducing the number of scanning angles, which is of particular interest in minimizing the measurement time.

3D Reconstruction of Optical Diffraction Tomography Based on a Neural Network Model

Morteza Hasani Shoreh

Optical tomography has been widely investigated for biomedical imaging applications. In recent years, it has been combined with digital holography and has been employed to produce high quality images of phase objects such as cells. In this Thesis, we look into some of the newest optical Diffraction Tomography (DT) based techniques to solve Three-Dimensional (3D) reconstruction problems and discuss and compare some of the leading ideas and papers. Then we propose a neural-network-based algorithm to solve this problem and apply it on both synthetic and biological samples. Conventional phase tomography with coherent light and off axis recording is performed. The Beam Propagation Method (BPM) is used to model scattering and each x-y plane is modeled by a layer of neurons in the BPM. The network's output (simulated data) is compared to the experimental measurements and the error is used for correcting the weights of the neurons (the refractive indices of the nodes) using standard error back-propagation techniques. The proposed algorithm is detailed and investigated. Then, we look into resolution-conserving regularization and discuss a method for selecting regularizing parameters. In addition, the local minima and phase unwrapping problems are discussed and ways of avoiding them are investigated. It is shown that the proposed learning tomography (LT) achieves better performance than other techniques such as, DT especially when insufficient number or incomplete set of measurements is available. We also explore the role of regularization in obtaining higher fidelity images without losing resolution. It is experimentally shown that due to overcoming multiple scattering, the LT reconstruction greatly outperforms the DT when the sample contains two or more layers of cells or beads. Then, reconstruction using intensity measurements is investigated. 3D reconstruction of a live cell during apoptosis is presented in a time-lapse format. At the end, we present a final comparison with leading papers and commercially available systems. It is shown that -compared to other existing algorithms- the results of the proposed method have better quality. In particular, parasitic granular structures and the missing cone artifact are improved. Overall, the perspectives of our approach are pretty rich for high-resolution tomographic imaging in a range of practical applications.

EPFL2018

Selection and Aggregation of Ranking Criteria for Retrieval from Scientific Publication Databases

Martin Veselý

Selection and aggregation of ranking criteria became an important topic in information retrieval as search is getting more specialized and as volume of electronically available information grows. In this context, document ranking has undergone a shift from purely content-based ranking criteria to combined ranking schemes, integrating additional fea- tures, such as document popularity or impact, or link-based techniques that were widely applied after the success of the PageRank algorithm. In this thesis we experiment with selection and aggregation of ranking criteria, aiming at increasing performance of a specialized scientific search engine. Based on theoretical foundations across several research fields including social choice theory, information re- trieval and digital libraries, we focus on ranking in databases of scientific publications in the field of High Energy Physics, identifying criteria that are pertinent for ranking scientific publications and selecting and aggregating them within a unified framework. The first issue that we address is thus identification and selection of ranking criteria for scientific documents in High Energy Physics. The criteria include the traditional information retrieval relevance that is based on the word similarity, but also the document usage, citation counts and links. In this context we present a novel ranking criterion combining the Hirsch index and the download counts, that we call the d-Hirsch index, taking into account counts of document downloads and assigning the corresponding Hirsch index directly to a document. Criteria selection is then based on correlation analysis between ranked lists of documents. We propose that correlations of entire document listings should be replaced by measuring an overlap on the top-k of the resulting list that should better reflect the independence in terms of ranking. To this end we proposed a new measure for the overlap, the Mean Average Overlap (MeanO). The second issue that we address is the aggregation framework for ranked lists of documents, where we focus on applying linear combination and models trained with ma- chine learning techniques based on logistic regression. As individual scores that are used for ranking are not necessarily comparable to each other, we describe a unified model for normalizing ranking scores before their aggregation, based on statistical properties of the underlying ranking criteria. Another contribution of our work is related to creation of a referential of relevance judgments for information retrieval experimentation in databases of scientific publications in the domain of High Energy Physics. Until now there has been no such resource avail- able that would allow to carry out evaluation of specialized information retrieval in this domain. We propose a method for automated generation of referentials, assuming that document relevance is determined by the document usage. Our approach corresponds to a modification of the pooling method, where validating of relevance judgments is not done by experts in the domain, but is inferred from previous document usage as observed in past user behaviors. The developed ranking models, methods and algorithms were validated in the frame- work of the d-Rank prototype developed in collaboration between EPFL and CERN, aiming at the design of document retrieval mechanisms and interfaces combining a multi- tude of individual specialized ranking criteria. This prototype has been integrated within the CERN institutional repository of scientific publications. We carried out several experiments to support the thesis that rank aggregation pro- vides a superior performance for specialized information retrieval than ranking with single criteria. We experimented with the CERN document server document repository, as well as with artificially generated data showing that rank aggregation can contribute to increased performance of a specialized information retrieval system, as compared to a baseline of the best individually performing ranking criteria. We showed on an experiment done with artificially generated data that if scores of individual ranking criteria follow a comparable distribution skewed to the right, then ag- gregating such ranking criteria can lead to an improvement of the ranking performance in terms of the Mean Average Precision. We reproduced this result on an experiment carried out on a real-world data set created from the CERN database of scientific publications showing that aggregating freshness and the download frequency yielded better ranking results as compared to rankings done with the download frequency and freshness sepa- rately, based on the Mean Reciprocal Rank and S@10 evaluation measures. We observed that local aggregates obtained from documents that were part of the result set performed significantly better than global aggregates, obtained through involving all documents in a collection. These results were confirmed by a 10-fold randomized cross-validation test for both the Mean Reciprocal Rank and the S@10 evaluation measure. We compared performance of the d-Hirsch index with existing ranking criteria at the CERN document server and with aggregates obtained through linear combination of individual scores and through logistic regression. We did not observe significantly better ranking results when ranking documents with the d-Hirsch index or its aggregates as compared to other ranking criteria in the test set. We suggested that selecting ranking criteria based on a novel selection criterion, the mean average overlaps between ranked lists, provides a justified basis for criteria selection.

EPFL2012

Learning approaches to high-fidelity optical diffraction tomography

Joowon Lim

Optical diffraction tomography (ODT) provides us 3D refractive index (RI) distributions of transparent samples. Since RI values differ across different materials, they serve as endogenous contrasts. It, therefore, enables us to image without pre-processing of labeling which can disturb samples during measurement. It has been utilized in various applications to study hematology, morphological parameters, biochemical information, and so on. The fundamental principle of ODT reconstruction is to recover the 3D information from multiple 2D measurements. While we require 2D measurements acquired by fully scanning a sample, there exist missing measurements that we are not able to access due to the limited numerical apertures (NAs) in the optical system. This is called the missing cone problem since the parts which are not covered by the NAs form cone shapes. The missing cone problem degrades the final reconstruction by underestimating RI values and more severely elongating images along the optical axis. Another challenge in ODT reconstruction is to model the nonlinear relationship between a sample and the measurements. The first order of scattering is commonly considered while neglecting the other higher orders to linearize the relationship, however, this results in degradation of the final reconstruction as the higher orders of scattering become more pronounced. In this thesis, we aim at solving the challenges in ODT reconstruction to provide more accurate quantitative information, namely, RI distributions. The first approach is based on model-based iterative reconstruction schemes. We choose the beam propagation method (BPM) for the forward model in order to capture the high orders of scattering. Due to the similarity of the multi-layer structure of the BPM with that of neural networks used in deep learning, we call this scheme learning tomography (LT). We rigorously investigate the performance of LT over the conventional linear model-based reconstruction scheme. Furthermore, by applying a more advanced BPM for the forward model, we even improve the LT and demonstrate the dramatically improved performance by both simulations and experiments. The second approach is based on statistically learning artifacts present in final reconstructions using a deep neural network (DNN) from a large dataset. Unlike the previous approaches which require iterations, the DNN approach instantly reconstructs RI distributions. We demonstrate the use of DNN using red blood cells which are highly distorted by the missing cone problem. In order to overcome the lack of ground truth in 3D ODT reconstruction, we digitally generate a synthetic dataset. The reconstruction results from the network present highly accurate results for the synthetic test set. Most importantly, we obtain high-fidelity reconstructions of experimental data using the network trained only on the synthetic data. Unlike other imaging modalities, ODT provides 3D quantitative information without labeling. To fully benefit from the capacity of quantitative imaging, it is critical to solve the existing challenges in ODT reconstruction to produce high-fidelity reconstructions. In this contribution, we aim to resolve the major challenges in ODT reconstruction using various learning approaches, and we believe that it can further improve ODT as a powerful tool for various applications.

EPFL2020