Deep Learning and Entropy-Based Texture Features for Color Image Classification.

In the domain of computer vision, entropy-defined as a measure of irregularity-has been proposed as an effective method for analyzing the texture of images. Several studies have shown that, with specific parameter tuning, entropy-based approaches achieve high accuracy in terms of classification results for texture images, when associated with machine learning classifiers. However, few entropy measures have been extended to studying color images. Moreover, the literature is missing comparative analyses of entropy-based and modern deep learning-based classification methods for RGB color images. In order to address this matter, we first propose a new entropy-based measure for RGB images based on a multivariate approach. This multivariate approach is a bi-dimensional extension of the methods that have been successfully applied to multivariate signals (unidimensional data). Then, we compare the classification results of this new approach with those obtained from several deep learning methods. The entropy-based method for RGB image classification that we propose leads to promising results. In future studies, the measure could be extended to study other color spaces as well.

Colored Texture Analysis Fuzzy Entropy Methods with a Dermoscopic Application.

Texture analysis is a subject of intensive focus in research due to its significant role in the field of image processing. However, few studies focus on colored texture analysis and even fewer use information theory concepts. Entropy measures have been proven competent for gray scale images. However, to the best of our knowledge, there are no well-established entropy methods that deal with colored images yet. Therefore, we propose the recent colored bidimensional fuzzy entropy measure, FuzEnC2D, and introduce its new multi-channel approaches, FuzEnV2D and FuzEnM2D, for the analysis of colored images. We investigate their sensitivity to parameters and ability to identify images with different irregularity degrees, and therefore different textures. Moreover, we study their behavior with colored Brodatz images in different color spaces. After verifying the results with test images, we employ the three methods for analyzing dermoscopic images of malignant melanoma and benign melanocytic nevi. FuzEnC2D, FuzEnV2D, and FuzEnM2D illustrate a good differentiation ability between the two-similar in appearance-pigmented skin lesions. The results outperform those of a well-known texture analysis measure. Our work provides the first entropy measure studying colored images using both single and multi-channel approaches.

Parameter Analysis of Multiscale Two-Dimensional Fuzzy and Dispersion Entropy Measures Using Machine Learning Classification.

Two-dimensional fuzzy entropy, dispersion entropy, and their multiscale extensions (MFuzzyEn2D and MDispEn2D, respectively) have shown promising results for image classifications. However, these results rely on the selection of key parameters that may largely influence the entropy values obtained. Yet, the optimal choice for these parameters has not been studied thoroughly. We propose a study on the impact of these parameters in image classification. For this purpose, the entropy-based algorithms are applied to a variety of images from different datasets, each containing multiple image classes. Several parameter combinations are used to obtain the entropy values. These entropy values are then applied to a range of machine learning classifiers and the algorithm parameters are analyzed based on the classification results. By using specific parameters, we show that both MFuzzyEn2D and MDispEn2D approach state-of-the-art in terms of image classification for multiple image types. They lead to an average maximum accuracy of more than 95% for all the datasets tested. Moreover, MFuzzyEn2D results in a better classification performance than that extracted by MDispEn2D as a majority. Furthermore, the choice of classifier does not have a significant impact on the classification of the extracted features by both entropy algorithms. The results open new perspectives for these entropy-based measures in textural analysis.

Multiscale Entropy Analysis of Short Signals: The Robustness of Fuzzy Entropy-Based Variants Compared to Full-Length Long Signals.

Multiscale entropy (MSE) analysis is a fundamental approach to access the complexity of a time series by estimating its information creation over a range of temporal scales. However, MSE may not be accurate or valid for short time series. This is why previous studies applied different kinds of algorithm derivations to short-term time series. However, no study has systematically analyzed and compared their reliabilities. This study compares the MSE algorithm variations adapted to short time series on both human and rat heart rate variability (HRV) time series using long-term MSE as reference. The most used variations of MSE are studied: composite MSE (CMSE), refined composite MSE (RCMSE), modified MSE (MMSE), and their fuzzy versions. We also analyze the errors in MSE estimations for a range of incorporated fuzzy exponents. The results show that fuzzy MSE versions-as a function of time series length-present minimal errors compared to the non-fuzzy algorithms. The traditional multiscale entropy algorithm with fuzzy counting (MFE) has similar accuracy to alternative algorithms with better computing performance. For the best accuracy, the findings suggest different fuzzy exponents according to the time series length.

Painless local pressure application to test microvascular reactivity to ischemia.

BACKGROUND: Forearm cutaneous blood flux (CBF) measurement with post-occlusive reactive hyperemia (PORH) is uncomfortable and may not be devoid of risks. We aimed to investigate post-compression reactive hyperemia (PCRH) with a custom-made indenter that was designed to be easily used routinely by inexperienced observers. METHODS: Medical students evaluated PCRH with 1- to 4-min pressure applications of 16 to 34 kPa and PORH with 3-min forearm cuff occlusion using laser speckle contrast imaging in 15 healthy volunteers. Participants were asked to quantify their discomfort with a visual analogue scale (VAS) of 10 cm. Total ischemia (ISCH) was quantified by the product of CBF during ischemia and ischemia duration (min). We subtracted the CBF changes in the skin from a reference ipsilateral (PCRH) or contralateral (PORH) non-stimulated area. RESULTS: The average VAS was 1.0 for PCRH vs. 6.0 for PORH (p < 0.001). A strong linear relationship between ISCH and peak PCRH (r2 = 0.915, p < 0.001) was noted. Peak PORH values (63.9 laser perfusion units (LPU)) were significantly lower than all values of the 3-min PCRH (72.6 LPU), including the one obtained with 16 kPa. CONCLUSION: Inexperienced observers could test microvascular reactivity with PCRH without inducing the discomfort that is typically experienced with PORH. Further, PCRH elicits a higher peak response to ischemia compared with PORH. This extremely simple method could influence a broad spectrum of routine cutaneous microcirculation investigations, especially when a painful approach is particularly inadequate or if the patient is fragile. CLINICAL TRIAL REGISTRATION: NCT02861924.

(Multiscale) Cross-Entropy Methods: A Review.

Cross-entropy was introduced in 1996 to quantify the degree of asynchronism between two time series. In 2009, a multiscale cross-entropy measure was proposed to analyze the dynamical characteristics of the coupling behavior between two sequences on multiple scales. Since their introductions, many improvements and other methods have been developed. In this review we offer a state-of-the-art on cross-entropy measures and their multiscale approaches.

Centered and Averaged Fuzzy Entropy to Improve Fuzzy Entropy Precision.

Several entropy measures are now widely used to analyze real-world time series. Among them, we can cite approximate entropy, sample entropy and fuzzy entropy (FuzzyEn), the latter one being probably the most efficient among the three. However, FuzzyEn precision depends on the number of samples in the data under study. The longer the signal, the better it is. Nevertheless, long signals are often difficult to obtain in real applications. This is why we herein propose a new FuzzyEn that presents better precision than the standard FuzzyEn. This is performed by increasing the number of samples used in the computation of the entropy measure, without changing the length of the time series. Thus, for the comparisons of the patterns, the mean value is no longer a constraint. Moreover, translated patterns are not the only ones considered: reflected, inversed, and glide-reflected patterns are also taken into account. The new measure (so-called centered and averaged FuzzyEn) is applied to synthetic and biomedical signals. The results show that the centered and averaged FuzzyEn leads to more precise results than the standard FuzzyEn: the relative percentile range is reduced compared to the standard sample entropy and fuzzy entropy measures. The centered and averaged FuzzyEn could now be used in other applications to compare its performances to those of other already-existing entropy measures.

Monitoring microvascular perfusion variations with laser speckle contrast imaging using a view-based temporal template method.

PURPOSE: Laser speckle contrast imaging (LSCI) continues to gain an increased interest in clinical and research studies to monitor microvascular perfusion. Due to its high spatial and temporal resolutions, LSCI may lead to a large amount of data. The analysis of such data, as well as the determination of the regions where the perfusion varies, can become a lengthy and tedious task. We propose here to analyze if a view-based temporal template method, the motion history image (MHI) algorithm, may be of use in detecting the perfusion variations locations. METHODS: LSCI data recorded during three different kinds of perfusion variations are considered: (i) cerebral blood flow during spreading depolarization (SD) in a mouse; (ii) cerebral blood flow during SD in a rat; (iii) cerebral blood flow during cardiac arrest in a rat. Each of these recordings was processed with MHI. RESULTS: We show that, for the three pathophysiological situations, MHI identifies the area in which perfusion evolves with time. The results are more easily obtained compared with a visual inspection of all of the frames constituting the recordings. MHI also has the advantage of relying on a rather simple algorithm. CONCLUSIONS: MHI can be tested in clinical and research studies to aid the user in perfusion analyses.

Visualization of perfusion changes with laser speckle contrast imaging using the method of motion history image.

Laser speckle contrast imaging (LSCI) is a real-time imaging modality reflecting microvascular perfusion. We report on the application of the motion history image (MHI) method on LSCI data obtained from the two hemispheres of a mouse. Through the generation of a single image, MHI stresses the microvascular perfusion changes. Our experimental results performed during a pinprick-triggered spreading depolarization demonstrate the effectiveness of MHI: MHI allows the visualization of perfusion changes without loss of resolution and definition. Moreover, MHI provides close results to the ones given by the generalized differences (GD) algorithm. However, MHI has the advantage of giving information on the temporal evolution of the perfusion variations, which GD does not.

Microvascular blood flow monitoring with laser speckle contrast imaging using the generalized differences algorithm.

Laser speckle contrast imaging (LSCI) is a full-field optical technique to monitor microvascular blood flow with high spatial and temporal resolutions. It is used in many medical fields such as dermatology, vascular medicine, or neurosciences. However, LSCI leads to a large amount of data: image sampling frequency is often of several Hz and recordings usually last several minutes. Therefore, clinicians often perform regions of interest in which a spatial averaging of blood flow is performed and the result is followed with time. Unfortunately, this leads to a poor spatial resolution for the analyzed data. At the same time, a higher spatial resolution for the perfusion maps is wanted. To get over this dilemma we propose a new post-acquisition visual representation for LSCI perfusion data using the so-called generalized differences (GD) algorithm. From a stack of perfusion images, the procedure leads to a new single image with the same spatial resolution as the original images and this new image reflects perfusion changes. The algorithm is herein applied on simulated stacks of images and on experimental LSCI perfusion data acquired in three different situations with a commercialized laser speckle contrast imager. The results show that the GD algorithm provides a new way of visualizing LSCI perfusion data.

Assessment of endothelial function by acetylcholine iontophoresis: impact of inter-electrode distance and electrical cutaneous resistance.

OBJECTIVES: Endothelial function can be assessed by acetylcholine (ACh) iontophoresis with single current application. The effect of inter-electrode distance as well as electrical cutaneous resistance (ECR) on ACh dependent vasodilation has never been studied using single current application. The aims of this study are (i) to compare ACh-peak and ECR measured at different inter-electrode distances, (ii) to assess the relationship between ACh-peak and ECR, (iii) and to study the reproducibility of the ECR values. METHODS: Fourteen healthy subjects were included. Using laser speckle contrast imaging, ACh-iontophoreses (0.1 mA, 30s) were performed on the forearm at a 7-day interval with an inter-electrode distance set at 5 cm. Two other inter-electrode distances were also evaluated: 10 cm and 15 cm. ECR was measured during each ACh-iontophoresis as well as the ACh-peak. RESULTS: No statistical difference was found between the ACh-peak values obtained at 5 cm, 10 cm and 15 cm. ECRs were also not statistically different. An inverse relationship (r=-0.60) was found between the ACh-peak and ECR (p<0.05). The coefficient of variation of the inter-day reproducibility of the ECR values was 9.1% [6.5%-15.1%] with an intra-class-correlation coefficient of 0.93 [0.81-0.98]. CONCLUSION: Inter-electrode distance ranging from 5 cm to 15 cm changes neither the ACh-peak value nor the ECR value. ECR impacts ACh-peak values.

Effect of skin temperature on skin endothelial function assessment.

PURPOSE: Microcirculatory dysfunction plays a key role in the development of sepsis during which core temperature is often disturbed. Skin microvascular assessment using laser techniques has been suggested to evaluate microvascular dysfunction during sepsis, but skin microcirculation is also a major effector of human thermoregulation. Therefore we aimed to study the effect of skin temperature on endothelial- and non-endothelial microvascular responses. METHODS: Fifteen healthy participants were studied at different randomized ambient temperatures leading to low (28.0+/-2.0 °C), intermediate (31.6+/-2.1 °C), and high (34.1+/-1.3 °C) skin temperatures. We measured skin blood flow using laser speckle contrast imaging on the forearm in response to vasodilator microvascular tests: acetylcholine (ACh) iontophoresis, sodium nitroprussiate (SNP) iontophoresis, and post-occlusive reactive hyperemia (PORH). The results are expressed as absolute (laser speckle perfusion units, LSPU) or normalized values (cutaneous vascular conductance, CVC in LSPU/mmHg and multiple of baseline). RESULTS: Maximal vasodilation induced by these tests is modified by skin temperature. A low skin temperature induced a significant lower vasodilation for all microvascular tests when results are expressed either in absolute values or in CVC. For example, ACh peak was 57.6+/-19.6 LSPU, 66.8+/-22.2 LSPU and 88.5+/-13.0 LSPU for low, intermediate and high skin temperature respectively (p<0.05). When results are expressed in multiple of baseline, statistical difference disappeared. CONCLUSIONS: These results suggest that skin temperature has to be well controlled when performing microvascular assessments in order to avoid any bias. The effect of skin temperature can be corrected by expressing the results in multiple of baseline.

Pulmonary embolism detection on venous thrombosis ultrasound images with bi-dimensional entropy measures: Preliminary results.

BACKGROUND: Venous thromboembolism (VTE) is a common health issue. A clinical expression of VTE is a deep vein thrombosis (DVT) that may lead to pulmonary embolism (PE), a critical illness. When DVT is suspected, an ultrasound exam is performed. However, the characteristics of the clot observed on ultrasound images cannot be linked with the presence of PE. Computed tomography angiography is the gold standard to diagnose PE. Nevertheless, the latter technique is expensive and requires the use of contrast agents. PURPOSE: In this article, we present an image processing method based on ultrasound images to determine whether PE is associated or not with lower limb DVT. In terms of medical equipment, this new approach (Doppler ultrasound image processing) is inexpensive and quite easy. METHODS: With the aim to help medical doctors in detecting PE, we herein propose to process ultrasound images of patients with DVT. After a first step based on histogram equalization, the analysis procedure is based on the use of bi-dimensional entropy measures. Two different algorithms are tested: the bi-dimensional dispersion entropy ( D i s p E n 2 D $DispEn_{2D}$ ) mesure and the bi-dimensional fuzzy entropy ( F u z E n 2 D $FuzEn_{2D}$ ) mesure. Thirty-two patients (12 women and 20 men, 67.63 ± 16.19 years old), split into two groups (16 with and 16 without PE), compose our database of around 1490 ultrasound images (split into seven different sizes from 32× 32 px to 128 × 128 px). p-values, computed with the Mann-Whitney test, are used to determine if entropy values of the two groups are statistically significantly different. Receiver operating characteristic (ROC) curves are plotted and analyzed for the most significant cases to define if entropy values are able to discriminate the two groups. RESULTS: p-values show that there are statistical differences between F u z E n 2 D $FuzEn_{2D}$  of patients with PE and patients without PE for 112× 112 px and 128× 128 px images. Area under the ROC curve (AUC) is higher than 0.7 (threshold for a fair test) for 112× 112 and 128× 128 images. The best value of AUC (0.72) is obtained for 112× 112 px images. CONCLUSIONS: Bi-dimensional entropy measures applied to ultrasound images seem to offer encouraging perspectives for PE detection: our first experiment, on a small dataset, shows that F u z E n 2 D $FuzEn_{2D}$  on 112× 112 px images is able to detect PE. The next step of our work will consist in testing this approach on a larger dataset and in integrating F u z E n 2 D $FuzEn_{2D}$  in a machine learning algorithm. Furthermore, this study could also contribute to PE risk prediction for patients with VTE.

A self-attention model for cross-subject seizure detection.

Epilepsy is a neurological disorder characterized by recurring seizures, detected by electroencephalography (EEG). EEG signals can be detected by manual time-consuming analysis and recently by automatic detection. The latter poses a significant challenge due to the high dimensional and non-stationary nature of EEG signals. Recently, deep learning (DL) techniques have emerged as valuable tools for seizure detection. In this study, a novel data-driven model based on DL, incorporating a self-attention mechanism (SAT), is proposed. One notable advantage of the proposed method is its simplicity in application, as the raw signal data is directly fed into the suggested network without requiring expertise in signal processing. The model leverages a one-dimensional convolutional neural network (CNN) to extract relevant features from EEG signals. These features are then passed through a long short-term memory (LSTM) module to benefit from its memory capabilities, along with a SAT mechanism. The key contribution of this paper lies in the addition of the SAT layer to the LSTM encoder, enabling enhanced exploration of the latent mapping during the encoding step. Cross-subject experiments revealed good performance of this approach with F1-score of 97.8% and 92.7% for binary and five-class epileptic seizure recognition tasks, respectively, on the public UCI dataset, and 97.9% on the CHB-MIT database, surpassing state-of-the-art DL performance. Besides, the proposed method exhibits robustness to inter-subject variability.

Bidimensional ensemble entropy: Concepts and application to emphysema lung computerized tomography scans.

BACKGROUND AND OBJECTIVE: Bidimensional entropy algorithms provide meaningful quantitative information on image textures. These algorithms have the advantage of relying on well-known one-dimensional entropy measures dedicated to the analysis of time series. However, uni- and bidimensional algorithms require the adjustment of some parameters that influence the obtained results or even findings. To address this, ensemble entropy techniques have recently emerged as a solution for signal analysis, offering greater stability and reduced bias in data patterns during entropy estimation. However, such algorithms have not yet been extended to their two-dimensional forms. METHODS: We therefore propose six bidimensional algorithms, namely ensemble sample entropy, ensemble permutation entropy, ensemble dispersion entropy, ensemble distribution entropy, and two versions of ensemble fuzzy entropy based on different models or parameters initialization of an entropy algorithm. These new measures are first tested on synthetic images and further applied to a biomedical dataset. RESULTS: The results suggest that ensemble techniques are able to detect different levels of image dynamics and their degrees of randomness. These methods lead to more stable entropy values (lower coefficients of variations) for the synthetic data. The results also show that these new measures can obtain up to 92.7% accuracy and 88.4% sensitivity when classifying patients with pulmonary emphysema through a k-nearest neighbors algorithm. CONCLUSIONS: This is a further step towards the potential clinical deployment of bidimensional ensemble approaches to detect different levels of image dynamics and their successful performance on emphysema lung computerized tomography scans. These bidimensional ensemble entropy algorithms have potential to be used in various imaging applications thanks to their ability to distinguish more stable and less biased image patterns compared to their original counterparts.

Assessment and Comparison of Nonlinear Measures in Resting-State Magnetoencephalograms in Alzheimer's Disease and Mild Cognitive Impairment.

BACKGROUND: Nonlinear dynamical measures, such as fractal dimension (FD), entropy, and Lempel-Ziv complexity (LZC), have been extensively investigated individually for detecting information content in magnetoencephalograms (MEGs) from patients with Alzheimer's disease (AD). OBJECTIVE: To compare systematically the performance of twenty conventional and recently introduced nonlinear dynamical measures in studying AD versus mild cognitive impairment (MCI) and healthy control (HC) subjects using MEG. METHODS: We compared twenty nonlinear measures to distinguish MEG recordings from 36 AD (mean age = 74.06±6.95 years), 18 MCI (mean age = 74.89±5.57 years), and 26 HC subjects (mean age = 71.77±6.38 years) in different brain regions and also evaluated the effect of the length of MEG epochs on their performance. We also studied the correlation between these measures and cognitive performance based on the Mini-Mental State Examination (MMSE). RESULTS: The results obtained by LZC, zero-crossing rate (ZCR), FD, and dispersion entropy (DispEn) measures showed significant differences among the three groups. There was no significant difference between HC and MCI. The highest Hedge's g effect sizes for HC versus AD and MCI versus AD were respectively obtained by Higuchi's FD (HFD) and fuzzy DispEn (FuzDispEn) in the whole brain and was most prominent in left lateral. The results obtained by HFD and FuzDispEn had a significant correlation with the MMSE scores. DispEn-based techniques, LZC, and ZCR, compared with HFD, were less sensitive to epoch length in distinguishing HC form AD. CONCLUSIONS: FuzDispEn was the most consistent technique to distinguish MEG dynamical patterns in AD compared with HC and MCI.

Impact of experimental conditions on noncontact laser recordings in microvascular studies.

Microcirculation, especially skin microcirculation, is a window toward systemic vascular function in magnitude and underlying mechanisms. Different techniques have been developed to assess the microcirculation. Among these techniques, laser technology is used to perform noninvasive microvascular assessments. In the 1970s, the laser Doppler flowmetry (LDF) technique was proposed to monitor microvascular blood flow. More recently, noncontact technologies including laser Doppler perfusion imaging (LDI) and laser speckle contrast imaging (LSCI) have improved the reproducibility of the microcirculation measurements and facilitated some clinical evaluations such as on wounds and ulcers. However, due to the absence of contact between tissue and sensors, it is likely that different technical and environmental conditions may interfere with microvascular recordings. This review presents major technical and environmental conditions, which may interfere with noncontact laser recordings in microvascular studies.

Laser speckle contrast imaging: multifractal analysis of data recorded in healthy subjects.

PURPOSE: The monitoring of microvascular blood flow can now be performed with laser speckle contrast imaging (LSCI), a new noninvasive laser-based technique. LSCI images have good spatial and temporal resolutions. Nevertheless, from now, few processing of these data have been performed to have a better knowledge on their properties. We herein propose a multifractal analysis of LSCI data recorded in the forearm of healthy subjects, based on the method from Halsey et al., one of the popular methods using the box-counting technique. METHODS: In laser speckle contrast image time sequences, we studied time evolution of pixel values, as well as time evolution of pixel values averaged in regions of interest (ROI) of different sizes. The results are compared with the ones obtained with single-point laser Doppler flowmetry (LDF) signals recorded simultaneously to LSCI images. RESULTS: Our work shows that, for the range of scales studied and with the method from Halsey et al., time evolution of pixel values present narrow multifractal spectra, reminding the ones of monofractal data. However, we observe that when LSCI pixel values are averaged in ROI large enough and followed with time, the multifractal spectra become larger and closer to the ones of LDF signals. CONCLUSIONS: Single pixels from laser speckle contrast images may not possess the same multifractal properties as LDF signals. These findings could now be compared with the ones obtained with other ranges of scales and with data recorded from pathological subjects.