Deep neural networks (DNNs) have recently been applied and used in many advanced and diverse tasks, such as medical diagnosis, automatic driving, etc. Due to the lack of transparency of the deep models, DNNs are often criticized for their prediction that cannot be explainable by human. In this paper, we propose a novel Morphological Fragmental Perturbation Pyramid (MFPP) method to solve the Explainable AI problem. In particular, we focus on the black-box scheme, which can identify the input area that is responsible for the output of the DNN without having to understand the internal architecture of the DNN. In the MFPP method, we divide the input image into multi-scale fragments and randomly mask out fragments as perturbation to generate a saliency map, which indicates the significance of each pixel for the prediction result of the black box model. Compared with the existing input sampling perturbation method, the pyramid structure fragment has proved to be more effective. It can better explore the morphological information of the input image to match its semantic information, and does not need any value inside the DNN. We qualitatively and quantitatively prove that MFPP meets and exceeds the performance of state-of-the-art (SOTA) black-box interpretation method on multiple DNN models and datasets.

This paper presents outdoor field experimental results in the 28-GHz frequency band of a 5G radio access system where base stations (BSs) are installed on buildings to deploy a 5G trial area in the Tokyo Odaiba waterfront area. This is an early field trial assessment of key 5G features such as BS and mobile station (MS) beam tracking, being deployed in dense urban areas. In the trial area, a variety of key 5G technologies and conditions are evaluated such as BS and MS massive multiple-input multiple-output (MIMO) antennas, beamforming and tracking, use of the high 28-GHz frequency, the wide frequency bandwidth of 400 MHz, intra-baseband unit (BBU) hand over (HO), and a coverage area in a realistic urban deployment with four different trial courses. Experimental results show that throughput exceeding 1 Gbps is achieved in the experiment area, and approximately 200 Mbps is achieved at 500 m away from the BS. We confirm that beam tracking and intra-BBU HO operate effectively while compensating for high path loss in the 28-GHz band to achieve a coverage area of 500 m. On the other hand, line of sight (LoS) and Non-LoS conditions are critical to 5G performance in the 28-GHz band, and we find that 5G connections may drop behind trees, buildings, and/or under a footbridge. The trial results show that the downlink throughput exceeds 50 Mbps in 85%, 50%, 70%, and 33% of the four different trial courses, respectively.

This paper presents outdoor field experimental results in the 28-GHz frequency band of a 5G radio access system where base stations (BSs) and a core network are installed on buildings to deploy a 5G trial area in the Tokyo Odaiba waterfront area as a 5G trial site. This is the first report of a 5G trial in a realistic urban deployment in an approximate 1 square kilometer area using a single 5G network. In this trial area, a variety of key 5G technologies are evaluated such as the peak data rate and throughput performance, 5G coverage area, BS and mobile station (MS) massive multiple-input multiple-output antennas for beamforming and tracking, intra- and inter-baseband unit handover employing the high frequency of 28 GHz, and the wide frequency bandwidth of 400 MHz in a realistic urban deployment. Results show that coverage is provided for 88% of the assessed drive route and the downlink throughput of 650 Mbps is achieved for 80% of the route. As indicated when buildings block the radio signal, line of sight (LoS) or non-LoS (NLoS) conditions are critical for signal propagation in the 28-GHz frequency band. To improve coverage in NLoS areas, a portable BS using a wireless backhaul system is installed in the trial area. With the portable BS, the corresponding coverage area is increased from 50% to 83% along the assessed drive route resulting in improved downlink throughput from 600 Mbps to 660 Mbps, and 250 Mbps to 280 Mbps in the uplink.

In this paper, we consider an OFDM(A) based millimeter wave (mmWave) system that is derived from a 5x scaling of the LTE numerology and incorporates analog beamforming-based antenna systems at both the access point (eNB) and user equipment (UE). We characterize the performance of such a system via a single- and multi-user link budget analysis as well as detailed system simulations. The sensitivity of performance to inter-site distance (ISD) and indoor vs. outdoor UEs is evaluated. A key observation is that while mmWave systems can achieve high throughputs, especially for dense deployments, outage is a concern for indoor UEs and non-line-of-sight (NLoS) channels. This motivates the need to build in robustness in mmWave access systems design. An analysis of the throughput impact of in-band backhaul for delivering efficient end-to-end service in ultra-dense mmWave networks is shown. Optimal resource partitioning approaches are derived that yield design guidelines for dense network deployments with in-band backhaul.

In this paper cluster validation criteria for piecewise constant image segmentation are proposed. All the criteria are based on the maximum a posteriori (MAP) principle and derived and implemented by four different, but related approaches. They are obtained by using Taylor expansions and three of them are derived by Bayesian predictive densities. The third and fourth criteria are implemented by the bootstrap technique, and their evaluations are, therefore, computationally more intensive than the evaluations of the first two. The proposed rules are compared by computer simulations with the widely used Akaike's information criterion (AIC) and the minimum description length (MDL) criteria.

A novel method for edge detection in vector images is proposed that does not require any prior knowledge of the imaged scenes. In the derivation, it is assumed that the observed vector images are realizations of spatially quasistationary processes, and that the vector observations are generated by parametric probability distribution functions of known form whose parameters are in general unknown. The method detects and estimates the edge locations using a criterion derived by Bayesian theory. It chooses the number of edges and their locations according to the maximum a posteriori probability (MAP) principle. We provide results that demonstrate its performance on synthesized and real images.

In this correspondence, the objective is to segment vector images, which are modeled as multivariate finite mixtures. The underlying images are characterized by Markov random fields (MRFs), and the applied segmentation procedure is based on the expectation-maximization (EM) technique. We propose an initialization procedure that does not require any prior information and yet provides excellent initial estimates for the EM method. The performance of the overall segmentation is demonstrated by segmentation of simulated one-dimensional (1D) and multidimensional magnetic resonance (MR) brain images.

A new approach for the processing of piecewise-constant signals is proposed. It is based on modeling the observed data as a sum of a random signal and noise. The random signal has a Gibbs distribution, and the noise is Gaussian. A MAP criterion is derived for joint estimation of the number of signal levels and reconstruction of the signal. The criterion comprises of three terms, one corresponding to the likelihood of the data and two to penalties. One penalty term penalizes for unnecessary transitions, and the other, for unnecessary levels. The method has been tested on synthesized data and applied to single ion channel recordings.

We propose an unsupervised vector image segmentation technique that combines the iterated conditional modes (ICM) procedure with an initialization scheme that requires minimal prior knowledge. As is well known, every iterative segmentation procedure needs initialization parameters, which are usually obtained from training data. In the absence of such data, the initialization becomes a critical step towards accurate segmentation because bad initializations can lead to poor performance. Our initialization scheme, referred to as tree structure (TS) initialization, represents a sequence of binary searches and is similar to a method for data compression in coding theory. The scheme does not require any a priori information or initial parameters, except for the number of classes, and therefore is completely data-driven. Computer simulations on multidimensional magnetic resonance (MR) brain images are provided to demonstrate the overall excellent performance of the proposed TS-ICM method.

In recent years, many image segmentation approaches have been based on Markov random fields (MRFs). The main assumption of the MRF approaches is that the class parameters are known or can be obtained from training data. In this paper the authors propose a novel method that relaxes this assumption and allows for simultaneous parameter estimation and vector image segmentation. The method is based on a tree structure (TS) algorithm which is combined with Besag's iterated conditional modes (ICM) procedure. The TS algorithm provides a mechanism for choosing initial cluster centers needed for initialization of the ICM. The authors' method has been tested on various one-dimensional (1-D) and multidimensional medical images and shows excellent performance. In this paper the authors also address the problem of cluster validation. They propose a new maximum a posteriori (MAP) criterion for determination of the number of classes and compare its performance to other approaches by computer simulations.