A Novel Approach to Detect Abnormal Chest X-rays of COVID-19 Patients Using Image Processing and Deep Learning

2021 ◽  
pp. 23-41
Author(s):  
Subhagata Chattopadhyay
Keyword(s):  
Mean Squared Error ◽  
Signal To Noise Ratio ◽  
Median Filter ◽  
Hessian Matrix ◽  
Region Of Interest ◽  
Computational Time ◽  
X Rays ◽  
Feed Forward Neural Network ◽  
Novel Approach ◽  
Input Dataset

The study proposes a novel approach to automate classifying Chest X-ray (CXR) images of COVID-19 positive patients. All acquired images have been pre-processed with Simple Median Filter (SMF) and Gaussian Filter (GF) with kernel size (5, 5). The better filter is then identified by comparing Mean Squared Error (MSE) and Peak Signal-to-Noise Ratio (PSNR) of denoised images. Canny's edge detection has been applied to find the Region of Interest (ROI) on denoised images. Eigenvalues [-2, 2] of the Hessian matrix (5 × 5) of the ROIs are then extracted, which constitutes the 'input' dataset to the Feed Forward Neural Network (FFNN) classifier, developed in this study. Eighty percent of the data is used for training the said network after 10-fold cross-validation and the performance of the network is tested with the remaining 20% of the data. Finally, validation has been made on another set of 'raw' normal and abnormal CXRs. Precision, Recall, Accuracy, and Computational time complexity (Big(O)) of the classifier are then estimated to examine its performance.

scholarly journals PERBANDINGAN METODE UNTUK PERBAIKAN KUALITAS CITRA MAMMOGRAM

2016 ◽  
Vol 7 (2) ◽  
pp. 657
Author(s):  
Hanifah Rahmi Fajrin
Keyword(s):  
Mean Square Error ◽  
Signal To Noise Ratio ◽  
Median Filter ◽  
Region Of Interest ◽  
Contourlet Transform ◽  
Mean Square ◽  
Signal To Noise ◽  
Noise Ratio

Kanker payudara merupakan pembunuh nomor satu pada wanita di seluruh dunia. Sejauh ini, deteksi dari kanker payudara hanya menggunakan mata telanjang dan berdasarkan jam terbang (pengalaman) dari dokter dan radiologis. Terdapat beberapa kelemahan dalam menganalisis mammogram payudara guna mendeteksi keberadaan kanker payudara. Hal ini bisa diakibatkan oleh sel kanker yang tertutup oleh noise, kontras citra yang rendah dan faktor manusiawi lainnya seperti : kelelahan, mood, dan lainnya. Untuk meminimalisir hal tersebut dibutuhkan suatu metode yang dapat membantu dokter dalam menganalisis citra mammogram payudara. Pada penelitian ini, dilakukan suatu proses yang bertujuan untuk meningkatkan kualitas mammogram agar lebih memudahkan dokter dalam mendiagnosis kelainan pada payudara. Citra yang diolah merupakan citra mammogram yang tidak dipangkas dan tidak disegmentasi pada bagian Region of Interest (ROI), melainkan keseluruhan citra payudara setelah dihilangkan background yang berlebihan. Tahapan pada proses peningkatan kuliatas citra mammogram ini (pra pengolahan) terdiri dari : menghilangkan label atau artifak yang ditemukan pada mammogram, meng-crop citra pada bagian payudara saja (menghilangkan background), memperbaiki kontras citra dengan membandingkan beberapa metode yaitu: CLAHE, Non Subsampled Contourlet Transform (NSCT), Contras Stretching (CS) dan selanjutnya dilakukan smoothing dengan menggunakan filter median. Kinerja dari setiap metode dihitung dengan mencari nilai Mean Square Error (MSE) dan Peak Signal to Noise Ratio (PSNR) citra. Dari nilai MSE dan PSNR yang didapatkan, ditemukan nilai MSE dan PSNR terbaik pada metode NSCT dengan nilai 50.20878 db 31.75332 db. Kata kunci: CLAHE, NSCT, CS, median filter.

Performance Analysis of Median Filter With Respect to Different Padding Methods in the Context of Removing Salt and Pepper Noise

2019 ◽  
Vol 8 (3) ◽  
pp. 6-9
Author(s):  
T. Sudha ◽  
P. Nagendra Kumar
Keyword(s):  
Mean Squared Error ◽  
Signal To Noise Ratio ◽  
Median Filter ◽  
Speckle Noise ◽  
Salt And Pepper Noise ◽  
Original Image ◽  
Signal To Noise ◽  
Squared Error ◽  
Noise Ratio ◽  
Salt And Pepper

Image Processing is one of the major areas of research. Images are often corrupted with different types of noise such as Gaussian noise, Poisson noise, Salt and Pepper noise, Speckle noise etc.The present work analyses the performance of the median filter with respect to different padding methods in the context of removing salt and pepper noise.Peak Signal-to-Noise ratio and Mean Squared Error have been considered as parameters for performance evaluation. The results obtained show thatthe Peak Signal-to-Noise Ratio and Mean Squared Error obtained between the original image and the filtered image obtained by applying median filter with symmetric padding method on the image corrupted with salt and pepper noise is same as the Peak Signal-to-Noise Ratio and Mean Squared Error obtained between the original image and the filtered image obtained by applying median filter with replicate padding method on the image corrupted with salt and pepper noise respectively.

scholarly journals Region of Interest Prediction using Segmentation

2020 ◽  
Vol 9 (5) ◽  
pp. 613-617
Keyword(s):  
Mean Squared Error ◽  
Signal To Noise Ratio ◽  
Similarity Index ◽  
Region Of Interest ◽  
Structural Similarity ◽  
Signal To Noise ◽  
Edge Preservation ◽  
Segmentation Methods ◽  
Noise Ratio

Segmentation separates an image into different sections badsed on the desire of the user. Segmentation will be carried out in an image, until the region of interest (ROI) of an object is extracted. Segmentation reliability predicts the progress of the various segmentation techniques. In this paper, various segmentation methods are proposed and quality of segmentation is verified by using quality metrics like Mean Squared Error (MSE),Signal to Noise Ratio (SNR), Peak- Signal to Noise Ratio (PSNR), Edge Preservation Index (EPI) and Structural Similarity Index Metric (SSIM).

scholarly journals A novel approach to medical image watermarking for tamper detection and recovery of region of interest using block compression and checksum

2018 ◽  
Vol 7 (4) ◽  
pp. 2137 ◽  
Author(s):  
Gangadhara Rao K ◽  
Chaitanya Konda
Keyword(s):  
Medical Image ◽  
Signal To Noise Ratio ◽  
Image Watermarking ◽  
Region Of Interest ◽  
Ultra Sound ◽  
Compression Technique ◽  
Public Network ◽  
Medical Image Watermarking ◽  
Novel Approach ◽  
Effective Use

Effective use of telecommunication and information technology in telemedicine increases the medical services to the patients who are from far away locations. The doctors provide these services by evaluating the patient details & scans like CT Scan, MRI and Ultra Sound. The patient information is exchanged between doctors and patients on a public network which is not safe. In medical image, specific regions are very important to diagnosis known as Region of Interest (ROI) and the rest of the regions are not of much importance known as Region of Non-Interest (RONI). Providing security to the ROI is an important issue hence medical image watermarking is used to transmit the medical images by embedding the ROI into RONI. At the destination, if tampering is found in ROI then recovery of ROI is possible by extracting the ROI from RONI. In the proposed method, the medical image is divided into three parts: BORDER, ROI and RONI. Further the ROI and RONI are divided into blocks and each ROI block is mapped to RONI block by applying division hash function. Lossless block compression technique is applied to each ROI block and embedded the compressed ROI block into mapped RONI block. To provide authenticity to ROI, checksum is calculated for ROI and embed this checksum in BORDER. Again checksum is calculated for each ROI block and placed in mapped RONI blocks. Whether ROI is tampered or not, is to be identified by extracting the checksum from BORDER and if it is tampered then recover the ROI by mapped RONI. The efficiency of the proposed algorithm is estimated by the performance measures mainly Peak Signal to Noise Ratio (PSNR). The proposed method gives good results on average 55 dB of PSNR compared to the previous methods [21] by efficiently compressing the ROI and by checking the authenticity.

Image Quality Enhancing by Efficient Histogram Equalization

2014 ◽  
Vol 2 (2) ◽  
pp. 47-58
Author(s):  
Ismail Sh. Baqer
Keyword(s):  
Image Quality ◽  
Contrast Enhancement ◽  
Mean Squared Error ◽  
Signal To Noise Ratio ◽  
Histogram Equalization ◽  
Gray Level ◽  
Signal To Noise ◽  
Squared Error ◽  
Noise Ratio

A two Level Image Quality enhancement is proposed in this paper. In the first level, Dualistic Sub-Image Histogram Equalization DSIHE method decomposes the original image into two sub-images based on median of original images. The second level deals with spikes shaped noise that may appear in the image after processing. We presents three methods of image enhancement GHE, LHE and proposed DSIHE that improve the visual quality of images. A comparative calculations is being carried out on above mentioned techniques to examine objective and subjective image quality parameters e.g. Peak Signal-to-Noise Ratio PSNR values, entropy H and mean squared error MSE to measure the quality of gray scale enhanced images. For handling gray-level images, convenient Histogram Equalization methods e.g. GHE and LHE tend to change the mean brightness of an image to middle level of the gray-level range limiting their appropriateness for contrast enhancement in consumer electronics such as TV monitors. The DSIHE methods seem to overcome this disadvantage as they tend to preserve both, the brightness and contrast enhancement. Experimental results show that the proposed technique gives better results in terms of Discrete Entropy, Signal to Noise ratio and Mean Squared Error values than the Global and Local histogram-based equalization methods

scholarly journals Marker for the pre-clinical development of bone substitute materials

2017 ◽  
Vol 3 (2) ◽  
pp. 711-715
Author(s):  
Michael de Wild ◽  
Simon Zimmermann ◽  
Marcel Obrecht ◽  
Michel Dard
Keyword(s):  
Bone Graft ◽  
Mechanical Stability ◽  
Clinical Performance ◽  
Ex Vivo ◽  
Region Of Interest ◽  
Defect Site ◽  
Novel Approach ◽  
Bone Substitute Materials ◽  
Clinical Experiments

AbstractThin mechanically stable Ti-cages have been developed for the in-vivo application as X-ray and histology markers for the optimized evaluation of pre-clinical performance of bone graft materials. A metallic frame defines the region of interest during histological investigations and supports the identification of the defect site. This standardization of the procedure enhances the quality of pre-clinical experiments. Different models of thin metallic frameworks were designed and produced out of titanium by additive manufacturing (Selective Laser Melting). The productibility, the mechanical stability, the handling and suitability of several frame geometries were tested during surgery in artificial and in ex-vivo bone before a series of cages was preclinically investigated in the female Göttingen minipigs model. With our novel approach, a flexible process was established that can be adapted to the requirements of any specific animal model and bone graft testing.

scholarly journals A Comparative Study of 3D UE Positioning in 5G New Radio with a Single Station

Sensors ◽  
2021 ◽  
Vol 21 (4) ◽  
pp. 1178
Author(s):  
Bo Sun ◽  
Bo Tan ◽  
Wenbo Wang ◽  
Elena Simona Lohan
Keyword(s):  
Mean Squared Error ◽  
Signal To Noise Ratio ◽  
Time Estimation ◽  
Satellite System ◽  
Base Station ◽  
Rectangular Array ◽  
Single Station ◽  
5G Network ◽  
Autonomous Machines ◽  
3D Positioning

The 5G network is considered as the essential underpinning infrastructure of manned and unmanned autonomous machines, such as drones and vehicles. Besides aiming to achieve reliable and low-latency wireless connectivity, positioning is another function provided by the 5G network to support the autonomous machines as the coexistence with the Global Navigation Satellite System (GNSS) is typically supported on smart 5G devices. This paper is a pilot study of using 5G uplink physical layer channel sounding reference signals (SRSs) for 3D user equipment (UE) positioning. The 3D positioning capability is backed by the uniform rectangular array (URA) on the base station and by the multiple subcarrier nature of the SRS. In this work, the subspace-based joint angle-time estimation and statistics-based expectation-maximization (EM) algorithms are investigated with the 3D signal manifold to prove the feasibility of using SRSs for 3D positioning. The positioning performance of both algorithms is evaluated by estimation of the root mean squared error (RMSE) versus the varying signal-to-noise-ratio (SNR), the bandwidth, the antenna array configuration, and multipath scenarios. The simulation results show that the uplink SRS works well for 3D UE positioning with a single base station, by providing a flexible resolution and accuracy for diverse application scenarios with the support of the phased array and signal estimation algorithms at the base station.

scholarly journals Computing 3D Phase-Type Holograms Based on Deep Learning Method

Photonics ◽  
2021 ◽  
Vol 8 (7) ◽  
pp. 280
Author(s):  
Huadong Zheng ◽  
Jianbin Hu ◽  
Chaojun Zhou ◽  
Xiaoxi Wang
Keyword(s):  
Deep Learning ◽  
Signal To Noise Ratio ◽  
Contrast Ratio ◽  
Angular Spectrum ◽  
Structural Similarity ◽  
Computational Time ◽  
Experimental Investigations ◽  
Training Strategy ◽  
Holographic Display ◽  
Phase Type

Computer holography is a technology that use a mathematical model of optical holography to generate digital holograms. It has wide and promising applications in various areas, especially holographic display. However, traditional computational algorithms for generation of phase-type holograms based on iterative optimization have a built-in tradeoff between the calculating speed and accuracy, which severely limits the performance of computational holograms in advanced applications. Recently, several deep learning based computational methods for generating holograms have gained more and more attention. In this paper, a convolutional neural network for generation of multi-plane holograms and its training strategy is proposed using a multi-plane iterative angular spectrum algorithm (ASM). The well-trained network indicates an excellent ability to generate phase-only holograms for multi-plane input images and to reconstruct correct images in the corresponding depth plane. Numerical simulations and optical reconstructions show that the accuracy of this method is almost the same with traditional iterative methods but the computational time decreases dramatically. The result images show a high quality through analysis of the image performance indicators, e.g., peak signal-to-noise ratio (PSNR), structural similarity (SSIM) and contrast ratio. Finally, the effectiveness of the proposed method is verified through experimental investigations.

scholarly journals From Femtoseconds to Hours–Measuring Dynamics over 18 Orders of Magnitude with Coherent X-rays

2021 ◽  
Vol 11 (13) ◽  
pp. 6179
Author(s):  
Felix Lehmkühler ◽  
Wojciech Roseker ◽  
Gerhard Grübel
Keyword(s):  
Time Scales ◽  
Free Electron Laser ◽  
Photon Correlation Spectroscopy ◽  
Signal To Noise Ratio ◽  
Coherent Scattering ◽  
Correlation Spectroscopy ◽  
X Rays ◽  
Photon Correlation ◽  
X Ray ◽  
Scattering Technique

X-ray photon correlation spectroscopy (XPCS) enables the study of sample dynamics between micrometer and atomic length scales. As a coherent scattering technique, it benefits from the increased brilliance of the next-generation synchrotron radiation and Free-Electron Laser (FEL) sources. In this article, we will introduce the XPCS concepts and review the latest developments of XPCS with special attention on the extension of accessible time scales to sub-μs and the application of XPCS at FELs. Furthermore, we will discuss future opportunities of XPCS and the related technique X-ray speckle visibility spectroscopy (XSVS) at new X-ray sources. Due to its particular signal-to-noise ratio, the time scales accessible by XPCS scale with the square of the coherent flux, allowing to dramatically extend its applications. This will soon enable studies over more than 18 orders of magnitude in time by XPCS and XSVS.

scholarly journals A large area diamond-based beam tagging hodoscope for ion therapy monitoring

2018 ◽  
Vol 170 ◽  
pp. 09005 ◽  
Author(s):  
M.-L. Gallin-Martel ◽  
L. Abbassi ◽  
A. Bes ◽  
G. Bosson ◽  
J. Collot ◽  
...  
Keyword(s):  
Detection Efficiency ◽  
Signal To Noise Ratio ◽  
Therapy Monitoring ◽  
Chemical Vapor ◽  
X Rays ◽  
Reaction Products ◽  
Large Area ◽  
Position Measurement ◽  
Set Up ◽  
Good Signal

The MoniDiam project is part of the French national collaboration CLaRyS (Contrôle en Ligne de l’hAdronthérapie par RaYonnements Secondaires) for on-line monitoring of hadron therapy. It relies on the imaging of nuclear reaction products that is related to the ion range. The goal here is to provide large area beam detectors with a high detection efficiency for carbon or proton beams giving time and position measurement at 100 MHz count rates (beam tagging hodoscope). High radiation hardness and intrinsic electronic properties make diamonds reliable and very fast detectors with a good signal to noise ratio. Commercial Chemical Vapor Deposited (CVD) poly-crystalline, heteroepitaxial and monocrystalline diamonds were studied. Their applicability as a particle detector was investigated using α and β radioactive sources, 95 MeV/u carbon ion beams at GANIL and 8.5 keV X-ray photon bunches from ESRF. This facility offers the unique capability of providing a focused (~1 μm) beam in bunches of 100 ps duration, with an almost uniform energy deposition in the irradiated detector volume, therefore mimicking the interaction of single ions. A signal rise time resolution ranging from 20 to 90 ps rms and an energy resolution of 7 to 9% were measured using diamonds with aluminum disk shaped surface metallization. This enabled us to conclude that polycrystalline CVD diamond detectors are good candidates for our beam tagging hodoscope development. Recently, double-side stripped metallized diamonds were tested using the XBIC (X Rays Beam Induced Current) set-up of the ID21 beamline at ESRF which permits us to evaluate the capability of diamond to be used as position sensitive detector. The final detector will consist in a mosaic arrangement of double-side stripped diamond sensors read out by a dedicated fast-integrated electronics of several hundreds of channels.

Sign in / Sign up

Export Citation Format

Share Document