Search

When developing machine learning models, image quality assessment (IQA) measures are a crucial component for evaluation. However, commonly used IQA measures have been primarily developed and optimized for natural images. In many specialized settings, such as medical images, this poses an often-overlooked problem regarding suitability. In previous studies, the IQA measure HaarPSI showed promising behavior for natural and medical images. HaarPSI is based on Haar wavelet representations and the framework allows optimization of two parameters. So far, these parameters have been aligned for natural images. Here, we optimize these parameters for two annotated medical data sets, a photoacoustic and a chest X-Ray data set. We observe that they are more sensitive to the parameter choices than the employed natural images, and on the other hand both medical data sets lead to similar parameter values when optimized. We denote the optimized setting, which improves the performance for the medical images notably, by HaarPSI$_{MED}$. The results suggest that adapting common IQA measures within their frameworks for medical images can provide a valuable, generalizable addition to the employment of more specific task-based measures., Comment: 5 pages, 3 figures, 2 tables

Developing imaging models capable of detecting pathologies from chest X-rays can be cost and time-prohibitive for large datasets as it requires supervision to attain state-of-the-art performance. Instead, labels extracted from radiology reports may serve as distant supervision since these are routinely generated as part of clinical practice. Despite their widespread use, current rule-based methods for label extraction rely on extensive rule sets that are limited in their robustness to syntactic variability. To alleviate these limitations, we introduce RadPert, a rule-based system that integrates an uncertainty-aware information schema with a streamlined set of rules, enhancing performance. Additionally, we have developed RadPrompt, a multi-turn prompting strategy that leverages RadPert to bolster the zero-shot predictive capabilities of large language models, achieving a statistically significant improvement in weighted average F1 score over GPT-4 Turbo. Most notably, RadPrompt surpasses both its underlying models, showcasing the synergistic potential of LLMs with rule-based models. We have evaluated our methods on two English Corpora: the MIMIC-CXR gold-standard test set and a gold-standard dataset collected from the Cambridge University Hospitals., Comment: Accepted at BioNLP, ACL 2024

Image quality assessment (IQA) is standard practice in the development stage of novel machine learning algorithms that operate on images. The most commonly used IQA measures have been developed and tested for natural images, but not in the medical setting. Reported inconsistencies arising in medical images are not surprising, as they have different properties than natural images. In this study, we test the applicability of common IQA measures for medical image data by comparing their assessment to manually rated chest X-ray (5 experts) and photoacoustic image data (2 experts). Moreover, we include supplementary studies on grayscale natural images and accelerated brain MRI data. The results of all experiments show a similar outcome in line with previous findings for medical images: PSNR and SSIM in the default setting are in the lower range of the result list and HaarPSI outperforms the other tested measures in the overall performance. Also among the top performers in our medical experiments are the full reference measures FSIM, LPIPS and MS-SSIM. Generally, the results on natural images yield considerably higher correlations, suggesting that additional employment of tailored IQA measures for medical imaging algorithms is needed.

Image quality assessment (IQA) is not just indispensable in clinical practice to ensure high standards, but also in the development stage of novel algorithms that operate on medical images with reference data. This paper provides a structured and comprehensive collection of examples where the two most common full reference (FR) image quality measures prove to be unsuitable for the assessment of novel algorithms using different kinds of medical images, including real-world MRI, CT, OCT, X-Ray, digital pathology and photoacoustic imaging data. In particular, the FR-IQA measures PSNR and SSIM are known and tested for working successfully in many natural imaging tasks, but discrepancies in medical scenarios have been noted in the literature. Inconsistencies arising in medical images are not surprising, as they have very different properties than natural images which have not been targeted nor tested in the development of the mentioned measures, and therefore might imply wrong judgement of novel methods for medical images. Therefore, improvement is urgently needed in particular in this era of AI to increase explainability, reproducibility and generalizability in machine learning for medical imaging and beyond. On top of the pitfalls we will provide ideas for future research as well as suggesting guidelines for the usage of FR-IQA measures applied to medical images.

We present a novel clustering algorithm, visClust, that is based on lower dimensional data representations and visual interpretation. Thereto, we design a transformation that allows the data to be represented by a binary integer array enabling the use of image processing methods to select a partition. Qualitative and quantitative analyses measured in accuracy and an adjusted Rand-Index show that the algorithm performs well while requiring low runtime and RAM. We compare the results to 6 state-of-the-art algorithms with available code, confirming the quality of visClust by superior performance in most experiments. Moreover, the algorithm asks for just one obligatory input parameter while allowing optimization via optional parameters. The code is made available on GitHub and straightforward to use., Comment: 28 pages

We photoionize nitrogen molecules with a train of extreme ultraviolet attosecond pulses together with a weak infrared field. We measure the phase of the two-color two-photon ionization transition (molecular phase) for different states of the ion. We observe a 0.76 $\pi$ shift for the electrons produced in the ionization channels leading to the X$^2 \Sigma_g^+$, $v'=1$ and $v'=2$ states. We relate this phase shift to the presence of a complex resonance in the continuum. By providing both a high spectral and temporal resolution, this general approach gives access to the evolution of extremely short lived states, which is hardly accessible otherwise.

Optical coherence tomography angiography (OCTA) is a novel noninvasive imaging modality for visualization of retinal blood flow in the human retina. Using specific OCTA imaging biomarkers for the identification of pathologies, automated image segmentations of the blood vessels can improve subsequent analysis and diagnosis. We present a novel segmentation method for vessel density identification based on frequency representations of the image, in particular, using so-called Gabor filter banks. The algorithm is evaluated qualitatively and quantitatively on an OCTA image in-house data set from $10$ eyes acquired by a Cirrus HD-OCT device. Qualitatively, the segmentation outcomes received very good visual evaluation feedback by experts. Quantitatively, we compared the resulting vessel density values with automated in-built values provided by the device. The results underline the visual evaluation. For the evaluation of the FAZ identification substep, manual annotations of $2$ expert graders were used, showing that our results coincide well in visual and quantitative manners. Lastly, we suggest the computation of adaptive local vessel density maps that allow straightforward analysis of retinal blood flow in a local manner.

MRI signal acquisition with multiple coils in a phased array is nowadays commonplace. The use of multiple receiver coils increases the signal-to-noise ratio (SNR) and enables accelerated parallel imaging methods. Some of these methods, like GRAPPA or SPIRiT, yield individual coil images in the k-space domain which need to be combined to form a final image. Coil combination is often the last step of the image reconstruction, where the root sum of squares (rSOS) is frequently used. This straightforward method works well for coil images with high SNR, but can yield problems in images with artifacts or low SNR in all individual coils. We aim to analyze the final coil combination step in the framework of linear compression, including principal component analysis (PCA). With two data sets, a simulated and an in-vivo, we use random projections as a representation of the whole space of orthogonal projections. This allows us to study the impact of linear compression in the image space with diverse measures of reconstruction accuracy. In particular, the $L_2$ error, variance, SNR, and visual results serve as performance measures to describe the final image quality. We study their relationships and observe that the $L_2$ error and variance strongly correlate, but as expected minimal $L_2$ error does not necessarily correspond to the best visual results. In terms of visual evaluation and SNR, the compression with PCA outperforms all other methods, including rSOS on the uncompressed image space data.

Segmenting anatomical structures such as the photoreceptor layer in retinal optical coherence tomography (OCT) scans is challenging in pathological scenarios. Supervised deep learning models trained with standard loss functions are usually able to characterize only the most common disease appeareance from a training set, resulting in suboptimal performance and poor generalization when dealing with unseen lesions. In this paper we propose to overcome this limitation by means of an augmented target loss function framework. We introduce a novel amplified-target loss that explicitly penalizes errors within the central area of the input images, based on the observation that most of the challenging disease appeareance is usually located in this area. We experimentally validated our approach using a data set with OCT scans of patients with macular diseases. We observe increased performance compared to the models that use only the standard losses. Our proposed loss function strongly supports the segmentation model to better distinguish photoreceptors in highly pathological scenarios., Comment: Accepted for publication at MICCAI-OMIA 2019

Convolutional neural network (CNN) architectures have originated and revolutionized machine learning for images. In order to take advantage of CNNs in predictive modeling with audio data, standard FFT-based signal processing methods are often applied to convert the raw audio waveforms into an image-like representations (e.g. spectrograms). Even though conventional images and spectrograms differ in their feature properties, this kind of pre-processing reduces the amount of training data necessary for successful training. In this contribution we investigate how input and target representations interplay with the amount of available training data in a music information retrieval setting. We compare the standard mel-spectrogram inputs with a newly proposed representation, called Mel scattering. Furthermore, we investigate the impact of additional target data representations by using an augmented target loss function which incorporates unused available information. We observe that all proposed methods outperform the standard mel-transform representation when using a limited data set and discuss their strengths and limitations. The source code for reproducibility of our experiments as well as intermediate results and model checkpoints are available in an online repository., Comment: 13 pages, 3 figures, 2 tables. Repository for reproducibility: https://gitlab.com/hararticles/gs-ms-mt/. Keywords: audio, CNN, limited data, Mel scattering, mel-spectrogram, augmented target loss function. Rewritten and restructured after peer revision. Recomputed and added new experiments and visualizations. Changed the presentation of the results

The use of orthogonal projections on high-dimensional input and target data in learning frameworks is studied. First, we investigate the relations between two standard objectives in dimension reduction, preservation of variance and of pairwise relative distances. Investigations of their asymptotic correlation as well as numerical experiments show that a projection does usually not satisfy both objectives at once. In a standard classification problem we determine projections on the input data that balance the objectives and compare subsequent results. Next, we extend our application of orthogonal projections to deep learning tasks and introduce a general framework of augmented target loss functions. These loss functions integrate additional information via transformations and projections of the target data. In two supervised learning problems, clinical image segmentation and music information classification, the application of our proposed augmented target loss functions increase the accuracy.

We briefly explain the use of cubature points on Grassmannians in numerical analysis.

Pulsators on and near the main sequence show period and amplitude changes that are too large to be the product of stellar evolution. The multiperiodic Delta Scuti stars are well suited to study these changes. This requires a very large amount of photometric data covering years and decades as well as mode identifications. We have examined over 800 nights of high-precision photometry of the multiperiodic pulsator 4 CVn obtained from 1966 through 2012. Because most of the data were obtained in adjacent observing seasons, it is possible to derive very accurate period values for a number of the excited pulsation modes and to study their systematic changes from 1974 to 2012. Most pulsation modes show systematic significant period and amplitude changes on a timescale of decades. For the well-studied modes, around 1986 a general reversal of the directions of both the positive and negative period changes occurred. Furthermore, the period changes between the different modes are strongly correlated, although they differ in size and sign. For the modes with known values of the spherical degree and azimuthal order, we find a correlation between the direction of the period changes and the identified azimuthal order, m. The associated amplitude changes generally have similar timescales of years or decades, but show little systematic or correlated behavior from mode to mode. A natural explanation for the opposite behavior of the prograde and retrograde modes is that their period changes are driven by a changing rotation profile. The changes in the rotation profile could in turn be driven by processes, perhaps the pulsations themselves, that redistribute angular momentum within the star. In general, different modes have different rotation kernels, so this will produce period shifts of varying magnitude for different modes.

Given a finite set of points on the Euclidean sphere, the worst case quadrature error in Sobolev spaces has recently been shown to provide upper bounds on the covering radius of the point set. Moreover, quasi-Monte Carlo integration points on the sphere achieve the asymptotically optimal covering radius. Here, we extend these results to points on compact smooth Riemannian manifolds and provide numerical experiments illustrating our findings for the Grassmannian manifold.

Numerical integration and function approximation on compact Riemannian manifolds based on eigenfunctions of the Laplace-Beltrami operator have been widely studied in the recent literature. The standard example in numerical experiments is the Euclidean sphere. Here, we derive numerically feasible expressions for the approximation schemes on the Grassmannian manifold, and we present the associated numerical experiments on the Grassmannian. Indeed, our experiments illustrate and match the corresponding theoretical results in the literature.

We present the results of a search for amplitude modulation of pulsation mode frequencies in 983 $\delta$ Sct stars, which have effective temperatures between 6400 $\leq T_{\rm eff} \leq$ 10 000 K in the Kepler Input Catalogue and were continuously observed by the Kepler Space Telescope for 4 yr. We demonstrate the diversity in pulsational behaviour observed, in particular nonlinearity, which is predicted for $\delta$ Sct stars. We analyse and discuss examples of $\delta$ Sct stars with constant amplitudes and phases; those that exhibit amplitude modulation caused by beating of close-frequency pulsation modes; those that exhibit pure amplitude modulation (with no associated phase variation); those that exhibit phase modulation caused by binarity; and those that exhibit amplitude modulation caused by nonlinearity. Using models and examples of individual stars, we demonstrate that observations of the changes in amplitude and phase of pulsation modes can be used to distinguish among the different scenarios. We find that 603 $\delta$ Sct stars (61.3 per cent) exhibit at least one pulsation mode that varies significantly in amplitude over 4 yr. Conversely, many $\delta$ Sct stars have constant pulsation amplitudes so short-length observations can be used to determine precise frequencies, amplitudes and phases for the most coherent and periodic $\delta$ Sct stars. It is shown that amplitude modulation is not restricted to a small region on the HR diagram, therefore not necessarily dependent on stellar parameters such as $T_{\rm eff}$ or $\log\,g$. Our catalogue of 983 $\delta$ Sct stars will be useful for comparisons to similar stars observed by K2 and TESS, because the length of the 4-yr Kepler data set will not be surpassed for some time., Comment: 21 pages, 12 figures, accepted for publication in MNRAS

The nature of period and amplitude changes in nonradial pulsators is presently unknown. It is therefore important to examine the correlations between these changes in stars with a large number of simultaneously excited pulsation modes. However, the small amplitudes require extensive high-precision photometry covering many years. We present 702 nights of high-precision photometry of the evolved Delta Scuti variable 4 CVn obtained from 2005 - 2012 with a dedicated telescope. We detected 64 frequencies, of which 38 can be identified as combinations and harmonics. The relative amplitudes of the combination frequencies are similar to those found in 44 Tau and show no evidence for resonant mode coupling. Significant period and amplitude changes are detected for the dominant modes. The known prograde and retrograde modes show period changes with opposite signs, while the radial mode exhibits only small, cyclical period changes. For each mode, the period changes are constant over the eight years and range from (1/P)dP/dt = -16 to +13 parts per million. On the other hand, the amplitude variations show no systematic behavior between different pulsation modes. The behavior of the prograde, axisymmetric, and retrograde modes indicates a constant decrease in the rotational splitting over the eight years.