key: cord-0464895-kzj9trp6 authors: Siracusano, Giulio; Corte, Aurelio La; Gaeta, Michele; Cicero, Giuseppe; Chiappini, Massimo; Finocchio, Giovanni title: Pipeline for Advanced Contrast Enhancement (PACE) of chest X-ray in evaluating COVID-19 patients by combining bidimensional empirical mode decomposition and CLAHE date: 2020-06-07 journal: nan DOI: nan sha: 31385df7e0b9fcd0bd7c3cdbbaa85eafebbcdf49 doc_id: 464895 cord_uid: kzj9trp6 COVID-19 is a new pulmonary disease which is driving stress to the hospitals due to the large number of cases worldwide. Imaging of lungs can play a key role in monitoring of the healthy status. Non-contrast chest computed tomography (CT) has been used for this purpose, mainly in China, with a significant success. However, this approach cannot be used massively mainly for both high risk and cost and in some countries also because this tool is not extensively available. Alternatively, chest X-ray, although less sensitive than CT-scan, can provide important information about the evolution of pulmonary involvement during the disease, this aspect is very important to verify the response of a patient to treatments. Here, we show how to improve the sensitivity of chest X-ray via a nonlinear post processing tool, named PACE, combining properly fast and adaptive bidimensional empirical mode decomposition and contrast limited adaptive histogram equalization (CLAHE). The results show an enhancement of the image contrast as confirmed by three widely used metrics: (i) contrast improvement index, (ii) entropy, and (iii) measure of enhancement. This improvement gives rise to a detectability of more lung lesions as identified by two radiologists, which evaluate the images separately, and confirmed by CT-scans. Based on our findings this method is proved as a flexible and effective way for medical image enhancement and can be used as a post-processing step for medical image understanding and analysis. Non-contrast chest Computed Tomography (CT) has many advantages over the planar chest X-ray which include the better spatial and densitometric resolution and the possibility of more clear identification of morphologic features of lesions. 1 Although chest CT can provide early diagnosis of COVID-19, it is not sufficient for that purpose considering that is aspecific, then for the World Health Organization the reference standard in the diagnosis of COVID-19 is the transcriptionpolymerase chain reaction which detects viral nucleic acid. 2 As already discussed in literature, chest CT can be used for the study of COVID-19 progression in patients providing quantitative information. 3 However, this approach needs several scans in a short window time. This fact gives rise to additional risks including increased possibility of cancer induction, because CT scan can expose a patient to as much radiations as 20-70 chest X-rays (CXR). 4 In addition, when used for evaluating COVID-19 patients, difficult and time consuming procedures of decontamination have to be set after each scan originating either increasing cost for patient and reducing availability time of the CT. Recent guidelines of North American Radiology Scientific Expert Panel have assessed that portable CXR has to be considered as the main imaging approach in evaluating COVID-19 patients 5 which not only reduces radiations for patients, but also avoids their transportation. The additional advantage of portable CXR is the possibility to monitor patients in the intensive care units (ICUs) which are more than 5% of the total known cases of COVID-19. Recent clinical studies 6, 7 confirm that CXRs can be used in describing some features in COVID-19 patients. The protocol to use CXR for evaluating COVID-19 patients has been set in the Hospital 'Policlinico G. Martino' of Messina (Italy) since the beginning of February 2020. In particular, an anteroposterior grey scale CXR in supine position has been acquired for each patient at bed with a portable X-ray equipment. Those images are often of low quality for the environment difficulties [8] [9] [10] and for noncollaborative and severely ill patients in most cases 11, 12 causing many different artifacts originating inhomogeneities in luminance distribution of radiograms. 13 To overcome these limitations that can impact on diagnostic effectiveness, here we develop a nonlinear post processing tool, we named Pipeline for Advanced Contrast Enhancement (PACE) which is aimed at improving the image quality of the CXR as evaluated in terms of contrast improvement index (CII), image entropy (ENT), and the measurement of enhancement (EME). FABEMD 17 is an algorithm, based on a variable window order-statistics filter for the calculation of the intrinsic mode functions (IMFs) in a bi-dimensional space (BIMFs). The filtering process iteratively separates and arranges the different BIMFs depending on the scale and fluctuations, from the faster to the slower ones where the residual image represents the latent information. Compared to other approaches to extract the BIMFs, FABEMD is fast because it does not need a recursive procedure. 18 In this work, the adopted FABEMD algorithm implementation can be summarized by the following steps. The j-th BIMF j  of an image i S is obtained by a neighboring window method 18 Step Step Both upper and lower envelopes are computed via the two parameters, which are defined as: max , , where , x y Z is the square region of window size, and , The residual image as computed via FABEMD is then filtered with the HMF   , H x y to the logtransformed image for correcting the non-uniform illumination of the CXR. 24 25 In particular here, the HMF is designed by using a high-pass filter to reduce or eliminate inhomogeneous effects originating by artifacts in CXR (gray images). 13 Here, we consider a modified version of the highpass filter   , HEF H x y based on High-Frequency Emphasis Filter (HEF) given by: The log-transformed image of pixel intensity of an image   , I x y ((x, y) are the spatial coordinates) can be expressed as a product between reflectance   , R x y of the object or scene and the intensity . 26 The   , L x y , which represents the image noise, can be reduced by its filtering in the log domain. 27, 28 Considering ( , ) Q x y as the logarithm of the pixel intensity ( , ) I x y : In the frequency domain, the Discrete Fourier Transform (DFT)  gives rise to the following expression: where Fig. 3 shows an example of the application of the HMF applied to the Residue achieved with the FABED algorithm as displayed in Fig. 2 (j). In summary, HMF represents the intermediate step in our pipeline and it has been implemented to integrate such filtering process with previous block for image segmentation (FABEMD) and the next reconstruction block which prepares the image for the final contrast enhancement by means of the CLAHE algorithm. The reconstruction block combines together the BIMFs with the filtered residual that will be then processed to generate a final single CXR with improved contrast, preserved details and where the illumination inhomogeneities are dramatically reduced. Fig. 3 (a) Residual image as obtained from FABEMD of the same input CXR image as processed in Fig. 2 (panel j). (b) Residual image after being processed using the HMF. As can be shown, the non-uniform luminance of the image has been significantly corrected (see red arrows as reference). The key difference between CLAHE and the Adaptive Histogram Equalization 32 is the approach to overcome the noise amplification problem. Basically, CLAHE splits the input image into nonoverlapping contextual regions (also called sub-images, tiles or blocks) and then applies histogram Nowadays, specific processing methods are available 39 to enhance image quality for specific applications. 40, 41 In order to assess the improvement, different enhancement metrics can be considered, such as contrast-based metrics computed in spatial domain. 42 Previous works have demonstrated that the combination of several enhancement techniques is required to improve the overall image quality. In this new method, the improvement is focused on homogeneous contrast, visibility, and image detail as they are the most essential and important factors that used in detection, recognition, and monitoring of lung lesions especially for COVID-19 patients. Following the results presented in literature, we have evaluated the performance of the proposed algorithm considering as reference metrics the CII, the ENT and EME, 43 which are examples of such spatial contrast measures. According to the literature, the CII is defined as follows 44 : where processed C and reference C are the contrast values of the processed and original image, respectively. The contrast C of a region is defined as 42 : p(k) being the probability to have a specific gray level value . The entropy, 45 which is a measure of the randomness characterizing the texture of the image, is estimated by the histogram of the image considered as a whole: where is the histogram count for an image segment. For an image   , x n m split into r c  blocks of size 1 2 I I  , the EME of is defined as 40 : The EME measure is suitable for images with attributes like noncomplex segments (e.g. regular geometrical shapes, like for human body parts), small targets in segments, non-periodic pattern in segments, and little to no randomness in segments. 46 In addition, many literatures on contrast enhancement 47-50 adopted EME as an evaluation indicator. for all the images based on all the three metrics have been obtained. In particular, an average increase of 10% in CII, 7.5% in Entropy, and 4.7% in EME have been documented. A visual example of the image improvement is shown in Fig. 4 (d-f) , where an CXR as acquired by the portable X-ray equipment (panel (d)), post-processed with CLAHE (panel (e)) and with the approach proposed here (panel (f)) are displayed, respectively. To better show the improvement, Fig. 4 (g-i) provide a magnified region for evaluating a specific part of the images. The performance of PACE have been also benchmarked in a public database available (https://github.com/ieee8023/covid-chestxray-dataset) 51 showing, an average increase of 9% in CII, 2.4% in ENT, and 2% in EME. In general, from a clinical point of view, the key achievement is that the number of lesions identified in the original image and the one post processed with CLAHE is the same, while PACE shows the capability to enhance the quality of the CXR enough in order to find more lesions when present. This result is confirmed in 8 patients, for the other cases while the number of lesions is the same for the three images (original, CLAHE, and our method) the velocity of lesions search has been improved as confirmed independently by a third radiologist. To support our claims, in those cases we have also performed the CT scans. Fig. 5 Here, it is possible to see a case where the lesion can be observed in the enhanced image and not in the original one. Lastly, the panels (d-f) provide a magnified area with evidence of how lungs are differently shown in original (d), CT image (e) and with our method (f). Medical imaging has a significant impact on medical applications, and since the quality of healthcare directly affects the quality of living of a patient, using the images for improving the performance of the medical specialists is an important issue. We have developed an automatic post processing tool, to enhance CXR images for the detection of lungs lesion. These results are generally applicable to image enhancement. On the other hand, the capability to support image monitoring of the healthy status of CODIV-19 patients has been tested in 79 patients. From a technical point of view, the improved performance of PACE approach has been confirmed by three well known metrics, (i) contrast improvement index, (ii) entropy, and (iii) measure of enhancement. On the basis of analysis of these metrics shows that this method preserves the input image details more accurately and gives processed image with better contrast enhancement and reduced brightness inhomogeneities. Concretely, we believe that this tool can support the clinical assistance of COVID-19 patients by enhancing the readability of CXR attained by portable X-ray equipment and in addition furnishing capability to monitor patients in intensive care units. The authors declare no conflict of interest. The ethical committee at the University Hospital of Messina does not require approval for a work on retrieved and anonymized data. For any information contact the secretary of Ethical Committee at the University Hospital of Messina at the contact information posted in its website (https://pre.polime.it/comitato_etico_interaziendale). Signal Process Signal Process Signal Process Nonlinear Digital Filters Int. Conf. Syst. Signals Image Process Graph. Gems Signal Process Digital Image Processing 2013 IEEE Int. Conf. Inf. Autom. (IEEE, 2013) This work was partially support by PETASPIN association and MARIS s.c.a.r.l. The data that support the plots within this paper and other findings of this study are available from the corresponding authors upon reasonable request.