Deep Learning Image Processing Enables 40% Faster Spinal MR Scans Which Match or Exceed Quality of Standard of Care

Abstract Objective This prospective multicenter multireader study evaluated the performance of 40% scan-time reduced spinal magnetic resonance imaging (MRI) reconstructed with deep learning (DL). Methods A total of 61 patients underwent standard of care (SOC) and accelerated (FAST) spine MRI. DL was used to enhance the accelerated set (FAST-DL). Three neuroradiologists were presented with paired side-by-side datasets (666 series). Datasets were blinded and randomized in sequence and left-right display order. Image features were preference rated. Structural similarity index (SSIM) and per pixel L1 was assessed for the image sets pre and post DL-enhancement as a quantitative assessment of image integrity impact. Results FAST-DL was qualitatively better than SOC for perceived signal-to-noise ratio (SNR) and artifacts and equivalent for other features. Quantitative SSIM was high, supporting the absence of image corruption by DL processing. Conclusion DL enables 40% spine MRI scan time reduction while maintaining diagnostic integrity and image quality with perceived benefits in SNR and artifact reduction, suggesting potential for clinical practice utility.

Download Full-text

ANALISIS SENSITIVITAS VIDEO MPEG-4 BERDASARKAN STRUKTUR FRAME PADA TRANSMISI DVB-T

Jurnal Ilmiah Informatika Komputer ◽

10.35760/ik.2020.v25i2.2691 ◽

2020 ◽

Vol 25 (2) ◽

pp. 86-97

Author(s):

Sandy Suryo Prayogo ◽

Tubagus Maulana Kusuma

Keyword(s):

Deep Learning ◽

Bit Error Rate ◽

Error Rate ◽

Signal To Noise Ratio ◽

Similarity Index ◽

Structural Similarity ◽

Signal To Noise ◽

Structural Similarity Index ◽

Noise Ratio

DVB merupakan standar transmisi televisi digital yang paling banyak digunakan saat ini. Unsur terpenting dari suatu proses transmisi adalah kualitas gambar dari video yang diterima setelah melalui proses transimisi tersebut. Banyak faktor yang dapat mempengaruhi kualitas dari suatu gambar, salah satunya adalah struktur frame dari video. Pada tulisan ini dilakukan pengujian sensitifitas video MPEG-4 berdasarkan struktur frame pada transmisi DVB-T. Pengujian dilakukan menggunakan simulasi matlab dan simulink. Digunakan juga ffmpeg untuk menyediakan format dan pengaturan video akan disimulasikan. Variabel yang diubah dari video adalah bitrate dan juga group-of-pictures (GOP), sedangkan variabel yang diubah dari transmisi DVB-T adalah signal-to-noise-ratio (SNR) pada kanal AWGN di antara pengirim (Tx) dan penerima (Rx). Hasil yang diperoleh dari percobaan berupa kualitas rata-rata gambar pada video yang diukur menggunakan metode pengukuran structural-similarity-index (SSIM). Dilakukan juga pengukuran terhadap jumlah bit-error-rate BER pada bitstream DVB-T. Percobaan yang dilakukan dapat menunjukkan seberapa besar sensitifitas bitrate dan GOP dari video pada transmisi DVB-T dengan kesimpulan semakin besar bitrate maka akan semakin buruk nilai kualitas gambarnya, dan semakin kecil nilai GOP maka akan semakin baik nilai kualitasnya. Penilitian diharapkan dapat dikembangkan menggunakan deep learning untuk memperoleh frame struktur yang tepat di kondisi-kondisi tertentu dalam proses transmisi televisi digital.

Download Full-text

AN EFFICIENT RIPPLET-BASED SHRINKAGE TECHNIQUE FOR MR IMAGE RESTORATION

Biomedical Engineering Applications Basis and Communications ◽

10.4015/s1016237214500744 ◽

2014 ◽

Vol 26 (06) ◽

pp. 1450074

Author(s):

A. Sumaiya Begum ◽

S. Poornachandra

Keyword(s):

Image Restoration ◽

Signal To Noise Ratio ◽

Similarity Index ◽

Structural Similarity ◽

High Energy ◽

Computational Time ◽

Ripplet Transform ◽

Anatomical Images ◽

Magnetic Resonance Imaging Mri ◽

Energy Compaction

In this paper a new ripplet-based shrinkage technique is used to suppress noise from Magnetic Resonance Imaging (MRI). The propitious properties of ripplet transform such as anisotropy, high directionality, good localization, and high-energy compaction make the proposed method efficient and feature preserving when compared to other transforms. Ripplet transform provides efficient representation of edges in images with a higher potential for image processing applications such as image restoration, compression, and de-noising. The proposed method implies a new nonlinear ripplet-based shrinkage technique to extract the spatial and frequency information from MRI corrupted by noise. The choice of this new shrinkage technique is due to its simplicity, versatility, and its efficiency in removing noise from homogenous regions and those regions with singularities, when compared to the existing filtering techniques. Experiments were conducted on several diffusion weighed images and anatomical images. The results show that the proposed de-noising technique shows competitive performance compared to the current state-of-art methods. Qualitative validation was performed based on several quality metrics and profound improvement over existing methods was obtained. Higher values of Peak Signal to Noise Ratio (PSNR), Correlation Coefficient (CC), mean structural similarity index (MSSIM), and lower values of Root Mean Square Error (RMSE) and computational time were obtained for the proposed ripplet-based shrinkage technique when compared to the existing ones.

Download Full-text

Multiscale Transform and Shrinkage Thresholding Techniques for Medical Image Denoising – Performance Evaluation

Cybernetics and Information Technologies ◽

10.2478/cait-2020-0033 ◽

2020 ◽

Vol 20 (3) ◽

pp. 130-146

Author(s):

S. Shajun Nisha ◽

S. P. Raja

Keyword(s):

Image Denoising ◽

Medical Image ◽

Performance Metrics ◽

Signal To Noise Ratio ◽

Similarity Index ◽

Structural Similarity ◽

Signal To Noise ◽

Noise Ratio ◽

Magnetic Resonance Imaging Mri ◽

Qualitative Performance

AbstractDue to sparsity and multiresolution properties, Mutiscale transforms are gaining popularity in the field of medical image denoising. This paper empirically evaluates different Mutiscale transform approaches such as Wavelet, Bandelet, Ridgelet, Contourlet, and Curvelet for image denoising. The image to be denoised first undergoes decomposition and then the thresholding is applied to its coefficients. This paper also deals with basic shrinkage thresholding techniques such Visushrink, Sureshrink, Neighshrink, Bayeshrink, Normalshrink and Neighsureshrink to determine the best one for image denoising. Experimental results on several test images were taken on Magnetic Resonance Imaging (MRI), X-RAY and Computed Tomography (CT). Qualitative performance metrics like Peak Signal to Noise Ratio (PSNR), Weighted Signal to Noise Ratio (WSNR), Structural Similarity Index (SSIM), and Correlation Coefficient (CC) were computed. The results shows that Contourlet based Medical image denoising methods are achieving significant improvement in association with Neighsureshrink thresholding technique.

Download Full-text

The Importance of Loss Functions for Increasing the Generalization Abilities of a Deep Learning-Based Next Frame Prediction Model for Traffic Scenes

Machine Learning and Knowledge Extraction ◽

10.3390/make2020006 ◽

2020 ◽

Vol 2 (2) ◽

pp. 78-98 ◽

Cited By ~ 1

Author(s):

Sandra Aigner ◽

Marco Körner

Keyword(s):

Deep Learning ◽

Prediction Model ◽

Short Term Memory ◽

Mean Squared Error ◽

Signal To Noise Ratio ◽

Similarity Index ◽

Structural Similarity ◽

Loss Functions ◽

Evaluation Metrics ◽

Training Dataset

This paper analyzes in detail how different loss functions influence the generalization abilities of a deep learning-based next frame prediction model for traffic scenes. Our prediction model is a convolutional long-short term memory (ConvLSTM) network that generates the pixel values of the next frame after having observed the raw pixel values of a sequence of four past frames. We trained the model with 21 combinations of seven loss terms using the Cityscapes Sequences dataset and an identical hyper-parameter setting. The loss terms range from pixel-error based terms to adversarial terms. To assess the generalization abilities of the resulting models, we generated predictions up to 20 time-steps into the future for four datasets of increasing visual distance to the training dataset—KITTI Tracking, BDD100K, UA-DETRAC, and KIT AIS Vehicles. All predicted frames were evaluated quantitatively with both traditional pixel-based evaluation metrics, that is, mean squared error (MSE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM), and recent, more advanced, feature-based evaluation metrics, that is, Fréchet inception distance (FID), and learned perceptual image patch similarity (LPIPS). The results show that solely by choosing a different combination of losses, we can boost the prediction performance on new datasets by up to 55%, and by up to 50% for long-term predictions.

Download Full-text

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

Photonics ◽

10.3390/photonics8070280 ◽

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.

Download Full-text

Joint Encryption and Compression-Based Watermarking Technique for Security of Digital Documents

ACM Transactions on Internet Technology ◽

10.1145/3414474 ◽

2021 ◽

Vol 21 (1) ◽

pp. 1-20

Author(s):

A. K. Singh ◽

S. Thakur ◽

Alireza Jolfaei ◽

Gautam Srivastava ◽

MD. Elhoseny ◽

...

Keyword(s):

Signal To Noise Ratio ◽

Similarity Index ◽

Structural Similarity ◽

Digital Data ◽

The Internet ◽

Discrete Wavelet ◽

Compression Performance ◽

Strong Compression ◽

Multiple Watermarks ◽

Value Decomposition

Recently, due to the increase in popularity of the Internet, the problem of digital data security over the Internet is increasing at a phenomenal rate. Watermarking is used for various notable applications to secure digital data from unauthorized individuals. To achieve this, in this article, we propose a joint encryption then-compression based watermarking technique for digital document security. This technique offers a tool for confidentiality, copyright protection, and strong compression performance of the system. The proposed method involves three major steps as follows: (1) embedding of multiple watermarks through non-sub-sampled contourlet transform, redundant discrete wavelet transform, and singular value decomposition; (2) encryption and compression via SHA-256 and Lempel Ziv Welch (LZW), respectively; and (3) extraction/recovery of multiple watermarks from the possibly distorted cover image. The performance estimations are carried out on various images at different attacks, and the efficiency of the system is determined in terms of peak signal-to-noise ratio (PSNR) and normalized correlation (NC), structural similarity index measure (SSIM), number of changing pixel rate (NPCR), unified averaged changed intensity (UACI), and compression ratio (CR). Furthermore, the comparative analysis of the proposed system with similar schemes indicates its superiority to them.

Download Full-text

Encryption Based Image Watermarking Algorithm in 2DWT-DCT Domains

Sensors ◽

10.3390/s21165540 ◽

2021 ◽

Vol 21 (16) ◽

pp. 5540

Author(s):

Nayeem Hasan ◽

Md Saiful Islam ◽

Wenyu Chen ◽

Muhammad Ashad Kabir ◽

Saad Al-Ahmadi

Keyword(s):

Signal To Noise Ratio ◽

Image Watermarking ◽

Similarity Index ◽

Structural Similarity ◽

Reconstructed Image ◽

Discrete Wavelet ◽

Signal To Noise ◽

Noise Ratio ◽

Watermarking Scheme

This paper proposes an encryption-based image watermarking scheme using a combination of second-level discrete wavelet transform (2DWT) and discrete cosine transform (DCT) with an auto extraction feature. The 2DWT has been selected based on the analysis of the trade-off between imperceptibility of the watermark and embedding capacity at various levels of decomposition. DCT operation is applied to the selected area to gather the image coefficients into a single vector using a zig-zig operation. We have utilized the same random bit sequence as the watermark and seed for the embedding zone coefficient. The quality of the reconstructed image was measured according to bit correction rate, peak signal-to-noise ratio (PSNR), and similarity index. Experimental results demonstrated that the proposed scheme is highly robust under different types of image-processing attacks. Several image attacks, e.g., JPEG compression, filtering, noise addition, cropping, sharpening, and bit-plane removal, were examined on watermarked images, and the results of our proposed method outstripped existing methods, especially in terms of the bit correction ratio (100%), which is a measure of bit restoration. The results were also highly satisfactory in terms of the quality of the reconstructed image, which demonstrated high imperceptibility in terms of peak signal-to-noise ratio (PSNR ≥ 40 dB) and structural similarity (SSIM ≥ 0.9) under different image attacks.

Download Full-text

Image Deblurring Using Multi-Stream Bottom-Top-Bottom Attention Network and Global Information-Based Fusion and Reconstruction Network

Sensors ◽

10.3390/s20133724 ◽

2020 ◽

Vol 20 (13) ◽

pp. 3724

Author(s):

Quan Zhou ◽

Mingyue Ding ◽

Xuming Zhang

Keyword(s):

Deep Learning ◽

Spatial Information ◽

Structural Similarity ◽

Image Deblurring ◽

Image Features ◽

Human Vision ◽

Global Information ◽

Attention Network ◽

Multi Scale ◽

Blind Image Deblurring

Image deblurring has been a challenging ill-posed problem in computer vision. Gaussian blur is a common model for image and signal degradation. The deep learning-based deblurring methods have attracted much attention due to their advantages over the traditional methods relying on hand-designed features. However, the existing deep learning-based deblurring techniques still cannot perform well in restoring the fine details and reconstructing the sharp edges. To address this issue, we have designed an effective end-to-end deep learning-based non-blind image deblurring algorithm. In the proposed method, a multi-stream bottom-top-bottom attention network (MBANet) with the encoder-to-decoder structure is designed to integrate low-level cues and high-level semantic information, which can facilitate extracting image features more effectively and improve the computational efficiency of the network. Moreover, the MBANet adopts a coarse-to-fine multi-scale strategy to process the input images to improve image deblurring performance. Furthermore, the global information-based fusion and reconstruction network is proposed to fuse multi-scale output maps to improve the global spatial information and recurrently refine the output deblurred image. The experiments were done on the public GoPro dataset and the realistic and dynamic scenes (REDS) dataset to evaluate the effectiveness and robustness of the proposed method. The experimental results show that the proposed method generally outperforms some traditional deburring methods and deep learning-based state-of-the-art deblurring methods such as scale-recurrent network (SRN) and denoising prior driven deep neural network (DPDNN) in terms of such quantitative indexes as peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) and human vision.

Download Full-text

Cell-filtering-based multi-scale Shannon–Cosine wavelet denoising method for locust slice images

International Journal of Wavelets Multiresolution and Information Processing ◽

10.1142/s0219691319500358 ◽

2019 ◽

Vol 17 (05) ◽

pp. 1950035 ◽

Cited By ~ 3

Author(s):

Shenghan Mei ◽

Xiaochun Liu ◽

Shuli Mei

Keyword(s):

Signal To Noise Ratio ◽

Gaussian White Noise ◽

Similarity Index ◽

Structural Similarity ◽

Wavelet Denoising ◽

Experimental Results ◽

Transform Method ◽

Denoising Method ◽

Multi Scale ◽

Piecewise Smoothness

The locust slice images have all the features such as strong self-similarity, piecewise smoothness and nonlinear texture structure. Multi-scale interpolation operator is an effective tool to describe such structures, but it cannot overcome the influence of noise on images. Therefore, this research designed the Shannon–Cosine wavelet which possesses all the excellent properties such as interpolation, smoothness, compact support and normalization, then constructing multi-scale wavelet interpolative operator, the operator can be applied to decompose and reconstruct the images adaptively. Combining the operator with the local filter operator (mean and median), a multi-scale Shannon–Cosine wavelet denoising algorithm based on cell filtering is constructed in this research. The algorithm overcomes the disadvantages of multi-scale interpolation wavelet, which is only suitable for describing smooth signals, and realizes multi-scale noise reduction of locust slice images. The experimental results show that the proposed method can keep all kinds of texture structures in the slice image of locust. In the experiments, the locust slice images with mixture noise of Gaussian and salt–pepper are taken as examples to compare the performances of the proposed method and other typical denoising methods. The experimental results show that the Peak Signal-To-Noise Ratio (PSNR) of the denoised images obtained by the proposed method is greater 27.3%, 24.6%, 2.94%, 22.9% than Weiner filter, wavelet transform method, median and average filtering, respectively; and the Structural Similarity Index (SSIM) for measuring image quality is greater 31.1%, 31.3%, 15.5%, 10.2% than other four methods, respectively. As the variance of Gaussian white noise increases from 0.02 to 0.1, the values of PSNR and SSIM obtained by the proposed method only decrease by 11.94% and 13.33%, respectively, which are much less than other 4 methods. This shows that the proposed method possesses stronger adaptability.

Download Full-text

A Novel Hierarchical Coding Progressive Transmission Method for WMSN Wildlife Images

Sensors ◽

10.3390/s19040946 ◽

2019 ◽

Vol 19 (4) ◽

pp. 946 ◽

Cited By ~ 2

Author(s):

Wenzhao Feng ◽

Chunhe Hu ◽

Yuan Wang ◽

Junguo Zhang ◽

Hao Yan

Keyword(s):

Signal To Noise Ratio ◽

Similarity Index ◽

Structural Similarity ◽

Transmission Efficiency ◽

Signal Interference ◽

Progressive Transmission ◽

Wireless Multimedia ◽

Transmission Method ◽

Hierarchical Coding ◽

Object Region

In the wild, wireless multimedia sensor network (WMSN) communication has limited bandwidth and the transmission of wildlife monitoring images always suffers signal interference, which is time-consuming, or sometimes even causes failure. Generally, only part of each wildlife image is valuable, therefore, if we could transmit the images according to the importance of the content, the above issues can be avoided. Inspired by the progressive transmission strategy, we propose a hierarchical coding progressive transmission method in this paper, which can transmit the saliency object region (i.e. the animal) and its background with different coding strategies and priorities. Specifically, we firstly construct a convolution neural network via the MobileNet model for the detection of the saliency object region and obtaining the mask on wildlife. Then, according to the importance of wavelet coefficients, set partitioned in hierarchical tree (SPIHT) lossless coding is utilized to transmit the saliency image which ensures the transmission accuracy of the wildlife region. After that, the background region left over is transmitted via the Embedded Zerotree Wavelets (EZW) lossy coding strategy, to improve the transmission efficiency. To verify the efficiency of our algorithm, a demonstration of the transmission of field-captured wildlife images is presented. Further, comparison of results with existing EZW and discrete cosine transform (DCT) algorithms shows that the proposed algorithm improves the peak signal to noise ratio (PSNR) and structural similarity index (SSIM) by 21.11%, 14.72% and 9.47%, 6.25%, respectively.

Download Full-text