A METAHEURISTIC APPROACH FOR TETROLET BASED MEDICAL IMAGE COMPRESSION

Over recent times, medical imaging plays a significant role in clinical practices. Storing and transferring the huge volume of images becomes complicated without an efficient image compression technique. This paper proposes a compression algorithm that uses a Haar based wavelet transform called Tetrolet transform, which reduces the noise on the input images and decomposes with a 4 x 4 blocks of equal squares called tetrominoes. It opts for a decomposing using optimal scheme for achieving the input image into a sparse representation which gives a much-detailed performance for texture and edge information better than wavelet transform. Set Partitioning in Hierarchical Trees (SPIHT) is used for encoding the significant coefficients to achieve efficient image compression. It has been investigated with various metaheuristic algorithms. Experimental results prove that the proposed method outperforms the other transform-based compression in terms of PSNR, CR, and Complexity. Also, the proposed method shows an improved result with another state of work.

Image Complexity Guided Network Compression for Biomedical Image Segmentation

Compression is a standard procedure for making convolutional neural networks (CNNs) adhere to some specific computing resource constraints. However, searching for a compressed architecture typically involves a series of time-consuming training/validation experiments to determine a good compromise between network size and performance accuracy. To address this, we propose an image complexity-guided network compression technique for biomedical image segmentation. Given any resource constraints, our framework utilizes data complexity and network architecture to quickly estimate a compressed model which does not require network training. Specifically, we map the dataset complexity to the target network accuracy degradation caused by compression. Such mapping enables us to predict the final accuracy for different network sizes, based on the computed dataset complexity. Thus, one may choose a solution that meets both the network size and segmentation accuracy requirements. Finally, the mapping is used to determine the convolutional layer-wise multiplicative factor for generating a compressed network. We conduct experiments using 5 datasets, employing 3 commonly-used CNN architectures for biomedical image segmentation as representative networks. Our proposed framework is shown to be effective for generating compressed segmentation networks, retaining up to ≈95% of the full-sized network segmentation accuracy, and at the same time, utilizing ≈32x fewer network trainable weights (average reduction) of the full-sized networks.

A Metaheuristic Approach for Tetrolet-Based Medical Image Compression

Over recent times, medical imaging plays a significant role in clinical practices. Storing and transferring the huge volume of images becomes complicated without an efficient image compression technique. This paper proposes a compression algorithm that uses a Haar based wavelet transform called Tetrolet transform, which reduces the noise on the input images and decomposes with a 4 x 4 blocks of equal squares called tetrominoes. It opts for a decomposing using optimal scheme for achieving the input image into a sparse representation which gives a much-detailed performance for texture and edge information better than wavelet transform. Set Partitioning in Hierarchical Trees (SPIHT) is used for encoding the significant coefficients to achieve efficient image compression. It has been investigated with various metaheuristic algorithms. Experimental results prove that the proposed method outperforms the other transform-based compression in terms of PSNR, CR, and Complexity. Also, the proposed method shows an improved result with another state of work.

Two-dimensional satellite image compression using compressive sensing

Compressive sensing is receiving a lot of attention from the image processing research community as a promising technique for image recovery from very few samples. The modality of compressive sensing technique is very useful in the applications where it is not feasible to acquire many samples. It is also prominently useful in satellite imaging applications since it drastically reduces the number of input samples thereby reducing the storage and communication bandwidth required to store and transmit the data into the ground station. In this paper, an interior point-based method is used to recover the entire satellite image from compressive sensing samples. The compression results obtained are compared with the compression results from conventional satellite image compression algorithms. The results demonstrate the increase in reconstruction accuracy as well as higher compression rate in case of compressive sensing-based compression technique.

Compressed SAR Interferometry in the Big Data Era

Modern Synthetic Aperture Radar (SAR) missions provide an unprecedented massive interferometric SAR (InSAR) time series. The processing of the Big InSAR Data is challenging for long-term monitoring. Indeed, as most deformation phenomena develop slowly, a strategy of a processing scheme can be worked on reduced volume data sets. This paper introduces a novel ComSAR algorithm based on a compression technique for reducing computational efforts while maintaining the performance robustly. The algorithm divides the massive data into many mini-stacks and then compresses them. The compressed estimator is close to the theoretical Cramer–Rao lower bound under a realistic C-band Sentinel-1 decorrelation scenario. Both persistent and distributed scatterers (PSDS) are exploited in the ComSAR algorithm. The ComSAR performance is validated via simulation and application to Sentinel-1 data to map land subsidence of the salt mine Vauvert area, France. The proposed ComSAR yields consistently better performance when compared with the state-of-the-art PSDS technique. We make our PSDS and ComSAR algorithms as an open-source TomoSAR package. To make it more practical, we exploit other open-source projects so that people can apply our PSDS and ComSAR methods for an end-to-end processing chain. To our knowledge, TomoSAR is the first public domain tool available to jointly handle PS and DS targets.

Ultrasonographic evaluation of deep vein thrombosis related to the central catheter in hemodialytic patients

Background Point of care ultrasound (PoCUS) is a useful tool for the early diagnosis of thrombosis related to the central venous catheter for dialysis (TR-CVCd). However, the application of PoCUS is still not common as a bedside imaging examination and TR-CVCd remains often underdiagnosed in the routine practice. The aim of this study was to investigate if a compression technique for the diagnosis of TR-CVCd blindly performed by PoCUS experts and medical students is accurate when compared to a Doppler study. Methods Two medical students without prior knowledge in PoCUS received a short theoretical–practical training to evaluate TR-CVCd of the internal jugular vein by means of the ultrasound compression technique. After the training phase, patients with central venous catheter for dialysis (CVCd) were evaluated by the students in a private hemodialysis clinic. The results were compared to those obtained on the same population by doctors with solid experience in PoCUS, using both the compression technique and the color Doppler. Results Eighty-one patients were eligible for the study and the prevalence of TR-CVCd diagnosed by Doppler was 28.4%. The compression technique performed by the students and by experts presented, respectively, a sensitivity of 59.2% (CI 51.6–66.8) vs 100% and a specificity of 89.6% (CI 84.9–94.3) vs 94.8% (CI 91.4–98.2). Conclusion The compression technique in the hands of PoCUS experts demonstrated high accuracy in the diagnosis of TR-CVCd and should represent a standard in the routine examination of dialytic patients. The training of PoCUS inexperienced students for the diagnosis of TR-CVCd is feasible but did not lead to a sufficient level of sensitivity.

HHO-Based Vector Quantization Technique for Biomedical Image Compression in Cloud Computing

In the present digital era, the exploitation of medical technologies and massive generation of medical data using different imaging modalities, adequate storage, management, and transmission of biomedical images necessitate image compression techniques. Vector quantization (VQ) is an effective image compression approach, and the widely employed VQ technique is Linde–Buzo–Gray (LBG), which generates local optimum codebooks for image compression. The codebook construction is treated as an optimization issue solved with utilization of metaheuristic optimization techniques. In this view, this paper designs an effective biomedical image compression technique in the cloud computing (CC) environment using Harris Hawks Optimization (HHO)-based LBG techniques. The HHO-LBG algorithm achieves a smooth transition among exploration as well as exploitation. To investigate the better performance of the HHO-LBG technique, an extensive set of simulations was carried out on benchmark biomedical images. The proposed HHO-LBG technique has accomplished promising results in terms of compression performance and reconstructed image quality.