Light field digital images are novel image modalities for capturing a sampled representation of the plenoptic function. A large amount of data is typically associated to a single sample of a scene, and data compressio...
详细信息
ISBN:
(纸本)9781728144962
Light field digital images are novel image modalities for capturing a sampled representation of the plenoptic function. A large amount of data is typically associated to a single sample of a scene, and data compression tools are required in order to develop systems and applications for light field communications. This paper presents the study of the performance of a convolutional neural network autoencoder as a tool for digital light field image compression. Testing conditions and a framework for the experimental evaluation are proposed for this study. Different encoders and coding conditions are taken into consideration, obtained results are reported and critically discussed.
Botnets are the powerful and effective way of performing malicious activities over the internet. Over the years, it has evolved into many forms. Earlier bots used static IP to communicate with their command and contro...
详细信息
ISBN:
(纸本)9781538670019
Botnets are the powerful and effective way of performing malicious activities over the internet. Over the years, it has evolved into many forms. Earlier bots used static IP to communicate with their command and control server. This method stopped working as soon as that specific IP was identified and blocked. These days, domain fluxing botnets are mostly in practice. The idea is, using Dynamically Generation Algorithm (DGA) to generate domains and use it to connect with C&C server. Numerous researches have been done to detect DGA botnets. These includes deriving features based on alphanumeric distribution of DGA domains and performing classification on it. Other studies include network logs analysis, time series analysis etc. Most of these domain classification works rely upon the features developed and may not work well if the botmaster decides to generate domain with completely new features. We are concerned with developing algorithm that is resilient to feature change that also work well for domain generated by completely new algorithm that was not seen before. We generated 16 bit representation of domains using autoencoder and classified it as benign or DGA generated using supervised learning(with neural net and SVM). To make it work with previously unseen algorithm, we tweaked our method with mean activation of 16-bit domain representation. This helped improve classification accuracy for completely new set of domain generation algorithm by up to 16%.
Proposed in 1991, Least Mean Square Error Reconstruction for self-organizing network, shortly Lmser, was a further development of the traditional auto-encoder (AE) by folding the architecture with respect to the centr...
详细信息
ISBN:
(数字)9783030362041
ISBN:
(纸本)9783030362041;9783030362034
Proposed in 1991, Least Mean Square Error Reconstruction for self-organizing network, shortly Lmser, was a further development of the traditional auto-encoder (AE) by folding the architecture with respect to the central coding layer and thus leading to the features of Duality in Connection Weight (DCW) and Duality in Paired Neurons (DPN), as well as jointly supervised and unsupervised learning which is called Duality in Supervision Paradigm (DSP). However, its advantages were only demonstrated in a one-hidden-layer implementation due to the lack of computing resources and big data at that time. In this paper, we revisit Lmser from the perspective of deep learning, develop Lmser network based on multiple fully-connected layers, and confirm several Lmser functions with experiments on image recognition, reconstruction, association recall, and so on. Experiments demonstrate that Lmser indeed works as indicated in the original paper, and it has promising performance in various applications.
This paper introduces an autoencoder structure to transfer the eye makeup from an arbitrary reference image to a source image realistically and faithfully using both synthetic paired data and unpaired data in a semi-s...
详细信息
ISBN:
(纸本)9781728150239
This paper introduces an autoencoder structure to transfer the eye makeup from an arbitrary reference image to a source image realistically and faithfully using both synthetic paired data and unpaired data in a semi-supervised way. Different from the image domain transfer problem, our framework only needs one domain entity and follows an "encoding-swap-decoding" process. Makeup transfer is achieved by decoding the base representation from a source image and makeup representation from a reference image. Moreover, our method allows users to control the makeup degree by tuning makeup weight. To the best of our knowledge, there is no public large makeup dataset to evaluate data-driven approaches. We have collected a dataset of non-makeup images and with-makeup images of various eye makeup styles. Experiments demonstrate the effectiveness of our method with the state-of-the-art methods both qualitatively and quantitatively.
This paper describes application of Artificial Intelligence using machine learning and deep learning at our laser diode module manufacturing facility. Implementing A.I. into data analysis and classification problems, ...
详细信息
ISBN:
(纸本)9781728133362
This paper describes application of Artificial Intelligence using machine learning and deep learning at our laser diode module manufacturing facility. Implementing A.I. into data analysis and classification problems, various benefits such as quality control, human work reduction and efficient usage of big data have been obtained.
In this paper, a novel framework for global diagnosis of autism spectrum disorder (ASD) using task-based functional MRI data is presented. A speech fMRI experiment is held to obtain local features related to the funct...
详细信息
ISBN:
(纸本)9781728123141
In this paper, a novel framework for global diagnosis of autism spectrum disorder (ASD) using task-based functional MRI data is presented. A speech fMRI experiment is held to obtain local features related to the functional activity of the brain. This study proposes both global diagnosis and local diagnosis by analyzing brain brainnetome atlas (BNT) which will lead to the first step of providing personalized medicine. The diagnosis pipeline consists of four steps on functional MRI volumes. The experimental results show that the global classification accuracy of our framework is about 75.8% and is much higher than other alternatives. Finally, comprehensive brain maps are provided for different individuals to indicate the degree of susceptibility of each brain area for autism, moving towards the idea of personalized medicine.
This paper presents the implementation of a Generative Adversarial Network (GAN) and Adversarial autoencoder (AAE) trained in an unsupervised manner using micro-Doppler (mD) spectrograms of human gait. Once the GAN ne...
详细信息
ISBN:
(纸本)9782874870576
This paper presents the implementation of a Generative Adversarial Network (GAN) and Adversarial autoencoder (AAE) trained in an unsupervised manner using micro-Doppler (mD) spectrograms of human gait. Once the GAN network was trained, the domain where micro-Doppler feature learning happens is inspected. This domain is then accessed by building the AAE and different network visualizations are shown. The benefits of unsupervised training are highlighted by investigating the self-learned spectrogram features, revealing the potential of unsupervised adversarial training techniques for mD spectrogram feature learning methods.
Classification, target detection, and compression are all important tasks in analyzing hyperspectral imagery (HSI). Because of the high dimensionality of HSI, it is often useful to identify low-dimensional representat...
详细信息
Classification, target detection, and compression are all important tasks in analyzing hyperspectral imagery (HSI). Because of the high dimensionality of HSI, it is often useful to identify low-dimensional representations of HSI data that can be used to make analysis tasks tractable. Traditional linear dimensionality reduction (DR) methods are not adequate due to the nonlinear distribution of HSI data. Many nonlinear DR methods, which are successful in the general data processing domain, such as Local Linear Embedding (LLE) [1], Isometric Feature Mapping (ISOMAP) [2] and Kernel Principal Components Analysis (KPCA) [3], run very slowly and require large amounts of of memory when applied to HSI. For example, applying KPCA to the 512×217 pixel, 204-band Salinas image using a modern desktop computer (AMD FX-6300 Six-Core Processor, 32 GB memory) requires more than 5 days of computing time and 28GB memory! In this thesis, we propose two different algorithms for significantly improving the computational efficiency of nonlinear DR without adversely affecting the performance of classification task: Simple Linear Iterative Clustering (SLIC) superpixels and semi-supervised deep autoencoder networks (SSDAN). SLIC is a very popular algorithm developed for computing superpixels in RGB images that can easily be extended to HSI. Each superpixel includes hundreds or thousands of pixels based on spatial and spectral similarities and is represented by the mean spectrum and spatial position of all of its component pixels. Since the number of superpixels is much smaller than the number of pixels in the image, they can be used as input for nonlinearDR, which significantly reduces the required computation time and memory versus providing all of the original pixels as input. After nonlinear DR is performed using superpixels as input, an interpolation step can be used to obtain the embedding of each original image pixel in the low dimensional space. To illustrate the power of using superpi
Explainable neural models have gained a lot of attention in recent years. However, conventional encoder–decoder models do not capture information regarding the importance of the involved latent variables and rely on ...
详细信息
Explainable neural models have gained a lot of attention in recent years. However, conventional encoder–decoder models do not capture information regarding the importance of the involved latent variables and rely on a heuristic a-priori specification of the dimensionality of the latent space or its selection based on multiple trainings. In this paper, we focus on the efficient structuring of the latent space of encoder–decoder approaches for explainable data reconstruction and compression. For this purpose, we leverage the concept of Shapley values to determine the contribution of the latent variables on the model’s output and rank them according to decreasing importance. As a result, a truncation of the latent dimensions to those that contribute the most to the overall reconstruction allows a trade-off between model compactness (i.e. dimensionality of the latent space) and representational power (i.e. reconstruction quality). In contrast to other recent autoencoder variants that incorporate a PCA-based ordering of the latent variables, our approach does not require time-consuming training processes and does not introduce additional weights. This makes our approach particularly valuable for compact representation and compression. We validate our approach at the examples of representing and compressing images as well as high-dimensional reflectance data.
Tens of millions of people live blind, and their number is ever increasing. Visual-to-auditory sensory substitution (SS) encompasses a family of cheap, generic solutions to assist the visually impaired by conveying vi...
详细信息
Tens of millions of people live blind, and their number is ever increasing. Visual-to-auditory sensory substitution (SS) encompasses a family of cheap, generic solutions to assist the visually impaired by conveying visual information through sound. The required SS training is lengthy: months of effort is necessary to reach a practical level of adaptation. There are two reasons for the tedious training process: the elongated substituting audio signal, and the disregard for the compressive characteristics of the human hearing system.
To overcome these obstacles, we developed a novel class of SS methods, by training deep recurrent autoencoders for image-to-sound conversion. We successfully trained deep learning models on different datasets to execute visual-to-auditory stimulus conversion. By constraining the visual space, we demonstrated the viability of shortened substituting audio signals, while proposing mechanisms, such as the integration of computational hearing models, to optimally convey visual features in the substituting stimulus as perceptually discernible auditory components. We tested our approach in two separate cases. In the first experiment, the author went blindfolded for 5 days, while performing SS training on hand posture discrimination. The second experiment assessed the accuracy of reaching movements towards objects on a table. In both test cases, above-chance-level accuracy was attained after a few hours of training.
Our novel SS architecture broadens the horizon of rehabilitation methods engineered for the visually impaired. Further improvements on the proposed model shall yield hastened rehabilitation of the blind and a wider adaptation of SS devices as a consequence.
暂无评论