Compressed Voxel-Based Mapping Using Unsupervised Learning

Region-based compression technique is particularly useful for radiological archiving system as it allows diagnostically important regions to be compressed with near lossless quality while the non-diagnostically important regions (NROI) to be compressed at lossy quality. In this paper, we present a region-based compression technique tailored for MRI brain scans. In the proposed technique termed as automated arbitrary PCA (AAPCA), an automatic segmentation based on brain symmetrical property is used to separate the ROI from the background. The arbitrary-shape ROI is then compressed by block-to-row PCA algorithm (BTRPCA) based on a factorization approach. The ROI is optimally compressed with lower compression rate while the NROI is compressed with higher compression rate. The proposed technique achieves satisfactory segmentation performance. The subjective and objective evaluation performed confirmed that the proposed technique achieves better performance metrics (PSNR and CoC) and higher overall compression rate. The experimental results also demonstrated that the proposed technique is more superior to various state-of-the-art compression methods.

A hybrid compression method for the NMR data based on window averaging and Discrete Cosine Transform

This paper presents a systematic study of the effects of hyperspectral pixel dimensionality reduction on the pixel classification task. We use five dimensionality reduction methods-PCA, KPCA, ICA, AE, and DAE-to compress 301-dimensional hyperspectral pixels. Compressed pixels are subsequently used to perform pixel classifications. Pixel classification accuracies together with compression method, compression rates, and reconstruction errors provide a new lens to study the suitability of a compression method for the task of pixel classification. We use three high-resolution hyperspectral image datasets, representing three common landscape types (i.e. urban, transitional suburban, and forests) collected by the Remote Sensing and Spatial Ecosystem Modeling laboratory of the University of Toronto. We found that PCA, KPCA, and ICA post greater signal reconstruction capability; however, when compression rates are more than 90% these methods show lower classification scores. AE and DAE methods post better classification accuracy at 95% compression rate, however their performance drops as compression rate approaches 97%. Our results suggest that both the compression method and the compression rate are important considerations when designing a hyperspectral pixel classification pipeline.

With the development of integrated circuit technology, the scale of Field Programmable Gate Array (FPGA) devices and the quantity of on-chip resources have been increased dramatically, as well as the size of configuration files, rendering the configuration process time-consuming. Especially, in the cloud-based multiple FPGAs co-designs, the transmission of large scale bitstream configuration files in public networks may suffer unpredictable delay, which causes circuits with high real-time requirements failed to be timely executed.To address this challenge, we divide a configuration file into three portions which have different data structures, and different compression methods are proposed to fit the data structure of each part to obtain a lower compression rate. For the portion which describes the FPGA on-chip resources, an LSTM and 1D-CNN based arithmetic coding compression method is introduced. In this way, the compression problem is converted to an estimation of a conditional probability distribution problem. Finally, to improve the prediction efficiency of the neural network, a cache based neural network inference acceleration method is proposed. Our approach is validated on the standard benchmark data sets and the average compression rate is 28.89%. Compared with seven different compression methods, our approach improves the compression rate by up to 39.45%.

Ego-motion estimation is a critical task for both monocular odometry and depth estimation, which are significant for navigation and scene perception in endoscopy. Most of existing ego-motion estimation frameworks only utilize a single modality of visual feature. In this work, a novel multimodal visual feature-based framework is proposed to perform real-time vision-based ego-motion estimation for visual odometry and depth estimation in endoscopic scenes. In the framework, to extract correlation information of two adjacent endoscopic frames, a channel attention-based module is proposed to integrate multidimensional features from the concatenated image and a feature interaction module is proposed to fuse scene features from two observations. Moreover, a novel pose decoder based on depthwise separable convolution is proposed to extract multiscale feature representation. Furthermore, a fully supervised pipeline and a self-supervised pipeline with the proposed framework are designed for monocular odometry and depth estimation, respectively. In the experiments, the proposed framework is compared with several advanced methods on five different datasets of various endoscopic scenes. Both quantitative and qualitative results demonstrate the pipeline with the proposed framework provides the most accurate monocular odometry and depth estimation in different kinds of endoscopic scenes.

A novel compression-based 2D-chaotic sine map for enhancing privacy and security of biometric identification systems

With the development of the pavement management system (PMS), the amount of the acquired and transmitted date is getting much larger [1][2]. Regard this, a compression method is needed for data analyzing. As there are lots of differences between the image of PMS and general image in the luminance distribution and the useful information region, general compression methods can not fit the special compression requirement of the PMS image. Although the PMS develops rapidly, the relevant image compression has little significant study. Based on Huffman coding and edge detection, this paper provides an efficient compression method, which can reserve all the useful information of the original image. The preliminary results show that the compression ration is higher than IS, which means it can work for the PMS.

Unsupervised learning from EEG data for epilepsy: A systematic literature review.

Systematic Review of Machine Learning in Diagnosis of Myeloproliferative Neoplasia

Nowadays, the image's compression presents major interest for storage and communication. This paper deals with DNA's particular images and compression based on frequency and time frequency Transform methods. Three compression methods realized by such Transforms have been studied. The first is based on the Discrete Cosine Transform (DCT), the second used the “Wavelet Transform” and the third is realized by the Curvelet Transform. We have investigated all the methods and adopted them on DNA's images (Chromosome I of C. Elegans). Then we evaluated their performances by calculating the Compression Rate (CR), the Peak Signal to Noise Ratio (PSNR) and also the Structural Similarity (SSIM) Index. We obtained promising results, especially with the Curvelet Transform that was used to improve our approach. CR, PSNR, SSIM reached 93.63, 97.0767 and 1, respectively.

Facies identification from basic well logs is essential for optimizing well operation or constructing reservoir model. The sequential nature of well log data makes the Hidden Markov Model (HMM) a suitable unsupervised machine learning approach for clustering. This study investigates the effectiveness of HMM combined with window filtering for unsupervised facies detection and compares its performance with the widely used K-Means Clustering (KMC) method. Synthetic gamma ray (GR), density (DEN), and neutron (NEU) logs, along with geologist-guided facies in a channel depositional environment, were used as input data. HMM identifies facies by modeling hidden geological states based on sequential well log data. Transition probabilities capture facies relationship across depths. A mode-based moving window filter smoothed the initial results by reducing fluctuations, while preserving facies boundaries. Validation was performed by comparing the results with reference facies using accuracy, adjusted rand index (ARI), normalized mutual information (NMI), and visual assessment. Both HMM and KMC clustering revealed thin non-geological facies, primarily at the facies boundaries. In the middle of facies, HMM mitigated thin facies more effectively than KMC. These thin facies arose from fluctuations in the well logs. Visually, HMM provided more geologically realistic results, despite KMC sometimes showed slightly better performance metrics. To address thin artificial facies, a moving window filter was applied. This slightly improved clustering performance metrics but significantly enhanced geological consistency by removing most artificial facies at boundaries or within facies, particularly in the smoothed HMM. Smoothed HMM achieved slightly better or comparable accuracy and NMI to smoothed KMC. However, minor thin facies persisted in both methods, more notably in KMC, due to significant well log fluctuations. The filter effectively smoothed the facies by reducing variations in thin facies across small depth intervals, ultimately enhancing the facies continuity. This improved facies distribution coherency with geological expectation, particularly for the smoothed HMM. HMM with post-processing plays a crucial role in identifying geologically consistent facies while maintaining or slightly improving clustering performance metrics. Unsupervised machine learning that incorporates sequential information is vital for facies clustering from well log data. The proposed workflow which integrates HMM with window filtering, enhances clustering outcomes by producing more geologically consistent result and more interpretable facies. This approach aids geoscientists in reservoir characterization and modeling, enabling more informed decision-making.

Range image registration and surface reconstruction have been traditionally considered as two independent problems where the latter relies on the results of the for-mer. This paper presents a new approach to surface recov-ery from range images where these two processes are inte-grated and performed in a common volumetric representa-tion. The volumetric representation contains both implicitly represented reconstructed surface as the signed distance field and corresponding matching information in the form of the gradient of the distance field. This allows both simul-taneous and incremental registration where matching com-plexity is linear with respect to the number of images. This improvement leads to incremental modeling from range im-age acquisition to surface reconstruction. It is shown that the approach is tolerant to initial registration errors as well as to measurement errors while keeping the details of the initial range images. The paper describes the formalism of the approach. Experimental results demonstrate perfor-mance advantages, tolerance to aforementioned types of er-rors and, as an application, filtering using redundant range data without loss of sharp details on the reconstructed sur-face. 1

Effect of Strain Rate on the Structure Properties of Silica Aerogel

Mobile robots can be easily equipped with numerous sensors which can aid in the tasks of localization and ego-motion estimation. Two such examples are Inertial Measurement Units (IMU), which provide a gravity vector via pitch and roll angular velocities, and wide-angle or panoramic imaging devices. As the number of powerful devices on a single robot increases, an important problem arises in how to fuse the information coming from multiple sources to obtain an accurate and efficient motion estimate. The IMU provides real-time readings which can be employed in orientation estimation, while in principle an Omnidirectional camera provides enough information to estimate the full rigid motion (up to translational scale). However, in addition to being computationally overwhelming, such an estimation is traditionally based on the sensitive search for feature correspondences between image frames. In this paper we present a novel algorithm that exploits information from an IMU to reduce the five parameter motion search to a three-parameter estimation. For this task we formulate a generalized Hough transform which processes image features directly to avoid searching for correspondences. The Hough space is computed rapidly by re-treating the transform as a convolution of spherical images.

The article explores the technical and theoretical aspects of machine learning (ML) in addressing the escalating complexities of cybersecurity threats in the digital age since the ever-growing rise in cybercrime has prompted users to utilize newer approaches to raise the bar on cybersecurity. Research considers the adoption of machine learning (ML) technology as a cornerstone of virtually any contemporary problem in cyber security, particularly processes and techniques involved in problem analysis, detection, attack prediction, and even behavioral profiling. Elaborated on how ML makes a better response compared to traditional methods like signature-based detection by explaining how real-time analysis of massive data becomes possible. An overview of the important features of supervised and unsupervised learning is provided in the context of anomaly detection and malicious activity recognition with a focus on Support Vector Machine and Isolation Forests algorithms as well as a detailed look at the LSTM model for phishing URL evolution analysis. Also, those algorithms have been highlighted from the technical implementation side: supervised learning with Support Vector Machines using Scikit-Learn to classify network traffic trained on features like IP addresses and ports, unsupervised learning with Isolation Forests for anomaly detection in multidimensional data, and deep learning with Long Short-Term Memory (LSTM) networks for phishing URL analysis. This paper investigates significantly important difficulties in carrying out ML algorithms, such as class imbalance, adversarial attacks, and lack of model transparency. Such techniques as SMOTE (Synthetic Minority Over-sampling Technique) are proposed for developing training datasets, whereas model adversarial training and robust optimization methods are suggested for defense against malicious model exploitation. Also, the role of explainability methods such as SHAP and LIME are emphasized to build the trust and acceptance of automated ML systems in cybersecurity. Identified research opportunities and suggested that further testing be done on improving model robustness and performance metrics in constrained environments.