Introduction: In this study the evaluation of a Platelet-based Maximum Penalized Likelihood Estimation (MPLE) for denoising SPECT images was performed and compared with other denoising methods such as Wavelets or Butterworth filtration. Platelet-based MPLE factorization as a multiscale decomposition approach has been already proposed for better edges and surfaces representation due to Poisson noise and inherent smoothness of this kind of images.Methods: We applied this approach on both simulated and real SPECT images. Monte Carlo simulations were generated with the SIMSET package to model the physical processes and instrumentation used in emission imaging. Cardiac, brain and NEMA phantom SPECT images were obtained using a single-head, Argus model SPECT system. The performance of this method has been evaluated both qualitatively and quantitatively with power spectrum, SNR and noise level measurements on simulated and real SPECT images.Results: For NEMA phantom images, the measured noise levels before (Mb) and after (Ma) denoising with Platelet-based MPLE approach were Mb=2.1732, Ma=0.1399. In patient study for 32 cardiac SPECT images, the difference between noise level and SNR before and after the approach were (Mb=3.7607, SNRb=9.7762, Ma=0.7374, SNRa=41.0848) respectively. Thus the Coefficient of variance (C.V) of SNR values for denoised images with this algorithm as compared with Butterworth filter, (145/33%) was found. For 32 brain SPECT images the Coefficient Variance of SNR values, (196/17%) was obtained.Conclusion: Our results shows that, Platelet-based MPLE is a useful method for denoising SPECT images considering better homogenous image, improvements in SNR, better radioactive uptake in target organ and reduction of interfering activity from background radiation in comparison to that of other conventional denoising methods.

In recent privacy has emerged as one of the major concerns of deep learning, since it requires huge amount of personal data. Medical Image Analysis is one of the prominent areas where sensitive data are shared to a third party service provider. In this paper, a secure deep learning scheme called Metamorphosed Learning (MPLE) is proposed to protect the privacy of images in medical image analysis. An augmented convolutional layer and image morphing are two main components of MPLE scheme. Data providers morph the images without privacy information using image morphing component. The human unrecognizable image is then delivered to the service providers who then apply deep learning algorithms on morphed data using augmented convolution layer without any performance penalty. MPLE provides sturdy security and privacy with optimal computational overhead. The proposed scheme is experimented using VGG-16 network on CIFAR dataset. The performance of MPLE is compared with similar works such as GAZELLE and MiniONN and found that the MPLE attracts very less computational and data transmission overhead. MPLE is also analyzed for various adversarial attack and realized that the success rate is as low as 7. 9 × 10-90. The efficiency of the proposed scheme is proved through experimental and performance analysis.