The objective of this paper is to develop an approach to objectively evaluate foreground detection algorithms. The purpose of the desired application is human detection in images sequences. This paper presents in a fi...
详细信息
The objective of this paper is to develop an approach to objectively evaluate foreground detection algorithms. The purpose of the desired application is human detection in images sequences. This paper presents in a first part the state of the art on foreground detection algorithms. In a second part, we use realistic synthetic videos to perform our evaluation. This kind of videos is very interesting for the evaluation study because it permits to evaluate the influences of some artifacts (changing light, video quality, shadow). We use a common metric to quantify the efficiency of the 11 selected algorithms. Time computation is also considered.
Computer-aided skin cancer detection systems built with deep neural networks yield overconfident predictions on out-of-distribution examples. Motivated by the importance of out-of-distribution detection in these syste...
详细信息
ISBN:
(数字)9781728193601
ISBN:
(纸本)9781728193618
Computer-aided skin cancer detection systems built with deep neural networks yield overconfident predictions on out-of-distribution examples. Motivated by the importance of out-of-distribution detection in these systems and the lack of relevant benchmarks targeted for skin cancer classification, we introduce a rich collection of out-of-distribution datasets - designed to comprehensively evaluate state-of-the-art out-of-distribution algorithms with skin cancer classifiers. In addition, we propose an adaptation in the Gram-Matrix algorithm for out-of-distribution detection that generally performs better and faster than the original algorithm for the considered skin cancer classification task. We also include a detailed discussion comparing the various state-of-the-art out-of-distribution detection algorithms and identify avenues for future research.
Composite hypothesis testing problems admit a class of solutions that are as optimal as possible, absent a cost function. These indomitable detectors comprise a superset of all optimal solutions. However, evaluating t...
详细信息
Composite hypothesis testing problems admit a class of solutions that are as optimal as possible, absent a cost function. These indomitable detectors comprise a superset of all optimal solutions. However, evaluating them with finite precision arithmetic can generate unstable answers. Here we devise a general method for deriving indomitables, using an extended fusion technique. Then we show how to reverse the process to generate arbitrarily accurate approximations to indomitables that are also stable.
Community Structure is one of the most relevant features of real world networks. Detecting such structures in large scale networks is a challenging task in scientific world. These are similar to clusters in which intr...
详细信息
ISBN:
(纸本)9781479968329
Community Structure is one of the most relevant features of real world networks. Detecting such structures in large scale networks is a challenging task in scientific world. These are similar to clusters in which intra cluster density is more than the inter cluster density. This paper reviews the prominent community detection algorithms that detect both disjoint and overlapped communities. These algorithms are experimented on benchmark dataset Zachary's karate club. Obtained number of communities is compared with the ground truth. The quality measures namely modularity and Normalized Mutual Information (NMI) are computed for all disjoint community detection algorithms. As a result of voluminous research done in this area the overlapped communities are come into the picture. Overlapped community means that a node in the network may be affiliated to more than one community. To test these algorithms Omega index is also included in this survey. After reviewing all these algorithms, this survey concludes that quality and scalability are the major issues in this area and also the measure used for detecting communities needs more computational power. So, one need to use either high performance computing framework with Graphical Processing Units (GPU) or Hadoop framework for distributed computing. Hence, this will balance the trade-off between running time and quality.
Recording systems for disk drives read and detect data from tracks consisting of "read" data sectors and "servo" sectors between read sectors. In this work, we (1) present a new asynchronous servo ...
详细信息
Recording systems for disk drives read and detect data from tracks consisting of "read" data sectors and "servo" sectors between read sectors. In this work, we (1) present a new asynchronous servo detection algorithm based on interpolation (2) quantitatively demonstrate improvements in performance of the asynchronous algorithm over the prior synchronous algorithm in the presence of moderate to high RI. Lastly, we propose to use an asynchronous algorithm, which dispenses with the baud rate timing loop and uses an asynchronously sampled ADC output for servo detection.
This paper develops several parallel algorithms for target detection in hyperspectral imagery, considered to be a crucial goal in many remote sensing applications. In order to illustrate parallel performance of the pr...
详细信息
This paper develops several parallel algorithms for target detection in hyperspectral imagery, considered to be a crucial goal in many remote sensing applications. In order to illustrate parallel performance of the proposed parallel algorithms, we consider a massively parallel Beowulf cluster at NASA's Goddard Space Flight Center. Experimental results, collected by the AVIRIS sensor over the World Trade Center, just five days after the terrorist attacks, indicate that commodity cluster computers can be used as a viable tool to increase computational performance of hyperspectral target detection applications.
This study aims to evaluate a variety of existing and novel fall detection algorithms, for a waist mounted accelerometer based system. algorithms were tested against a comprehensive data-set recorded from 10 young hea...
详细信息
This study aims to evaluate a variety of existing and novel fall detection algorithms, for a waist mounted accelerometer based system. algorithms were tested against a comprehensive data-set recorded from 10 young healthy subjects performing 240 falls and 120 activities of daily living and 10 elderly healthy subjects performing 240 scripted and 52.4 hours of continuous unscripted normal activities. Results show that using a simple algorithm employing Velocity+Impact+Posture can achieve a low false-positive rate of less than 1 FP/day* (0.94FPs/day*) with a sensitivity of 94.6% and a specificity of 100%. The algorithms were tested using unsupervised continuous activities performed by elderly subjects living in the community, which is the target environment for a fall detection device.
We present some results about the comparison of two families of anomaly detection algorithms, specific to hyperspectral images analysis, both based on local statistical Hypothesis Testing (HT). The study has involved ...
详细信息
We present some results about the comparison of two families of anomaly detection algorithms, specific to hyperspectral images analysis, both based on local statistical Hypothesis Testing (HT). The study has involved the RX and the Gauss-Markov Random Fields (GMRF) anomaly detectors. These algorithms share a quite similar sliding window strategy and the approach of statistical HT but they differ notably in the way data are modeled and the parameters estimated. An underlying main question is actually the trade-off between describing the background clutter with many parameters (case of RX), harder to locally estimate, and considering a model with fewer parameters (case of GMRF), thus easier to estimate from the local clutter only. The study includes the issue of reducing the spectral dimension with linear second order and higher order statistics based methods. As an output of the study, RX will be included in the next release of the Orfeo Toolbox (OTB), an open source remote sensing image processing library developed by the CNES.
Community detection is starting to become an inevitable task for data science community and like-minded analysts to understand the hidden insights and trends in large complex networks. Through this paper we would like...
详细信息
Community detection is starting to become an inevitable task for data science community and like-minded analysts to understand the hidden insights and trends in large complex networks. Through this paper we would like to understand and comparatively analyze the prevalent and recent community detection techniques like Minimum Cut Method, Hierarchical Clustering, Girvan Newman Algorithm, Modularity Maximization, Statistical Inference, Clique Percolation & Maximum Permanence Method based on their network relevance, adaptability and robustness to changes and finally their capacity of size and speed. Finally, we will review the realworld application of these algorithms.
To improve the performance of the channel decoding in receivers for digital mobile radio systems the detector should be able to deliver reliability information about the symbol decisions (soft-decision). This paper is...
详细信息
To improve the performance of the channel decoding in receivers for digital mobile radio systems the detector should be able to deliver reliability information about the symbol decisions (soft-decision). This paper is concerned with procedures which allow a soft-decision output at sequential detection algorithms. Simulation results obtained with these algorithms for the GSM system show the achievable gain over hard-decision detectors and allow a comparison to the more complex Soft-Output Viterbi-algorithm.< >
暂无评论