PyTorchKR Papers

최신 논문

총 5,575편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:1807.08784

Fast Vessel Segmentation and Tracking in Ultra High-Frequency Ultrasound Images

Ultra High Frequency Ultrasound (UHFUS) enables the visualization of highly deformable small and medium vessels in the hand. Intricate vessel-based measurements, such as intimal wall thickness and vessel wall compliance, require sub-millimeter vessel tracking between B-scans. Our fast GPU-based approach combines the advantages of local phase analysis, a distance-regularized level set, and an Extended Kalman Filter (EKF), to rapidly segment and track the deforming vessel contour. We validated on 35 UHFUS sequences of vessels in the hand, and we show the transferability of the approach to 5 more diverse datasets acquired by a traditional High Frequency Ultrasound (HFUS) machine. To the best of our knowledge, this is the first algorithm capable of rapidly segmenting and tracking deformable vessel contours in 2D UHFUS images. It is also the fastest and most accurate system for 2D HFUS images.

arXiv:1807.08271공식 코드 · Papers with Code

RGBiD-SLAM for Accurate Real-time Localisation and 3D Mapping

In this paper we present a complete SLAM system for RGB-D cameras, namely RGB-iD SLAM. The presented approach is a dense direct SLAM method with the main characteristic of working with the depth maps in inverse depth parametrisation for the routines of dense alignment or keyframe fusion. The system consists in 2 CPU threads working in parallel, which share the use of the GPU for dense alignment and keyframe fusion routines. The first thread is a front-end operating at frame rate, which processes every incoming frame from the RGB-D sensor to compute the incremental odometry and integrate it in a keyframe which is changed periodically following a covisibility-based strategy. The second thread is a back-end which receives keyframes from the front-end. This thread is in charge of segmenting the keyframes based on their structure, describing them using Bags of Words, trying to find potential loop closures with previous keyframes, and in such case perform pose-graph optimisation for trajectory correction. In addition, our system allows is able to compute the odometry both with unregistered and registered depth maps, allowing to use customised calibrations of the RGB-D sensor. As a consequence in the paper we also propose a detailed calibration pipeline to compute customised calibrations for particular RGB-D cameras. The experiments with our approach in the TUM RGB-D benchmark datasets show results superior in accuracy to the state-of-the-art in many of the sequences. The code has been made available on-line for research purposes https://github.com/dangut/RGBiD-SLAM.

arXiv:1807.07433

Real-Time Stereo Vision for Road Surface 3-D Reconstruction

Stereo vision techniques have been widely used in civil engineering to acquire 3-D road data. The two important factors of stereo vision are accuracy and speed. However, it is very challenging to achieve both of them simultaneously and therefore the main aim of developing a stereo vision system is to improve the trade-off between these two factors. In this paper, we present a real-time stereo vision system used for road surface 3-D reconstruction. The proposed system is developed from our previously published 3-D reconstruction algorithm where the perspective view of the target image is first transformed into the reference view, which not only increases the disparity accuracy but also improves the processing speed. Then, the correlation cost between each pair of blocks is computed and stored in two 3-D cost volumes. To adaptively aggregate the matching costs from neighbourhood systems, bilateral filtering is performed on the cost volumes. This greatly reduces the ambiguities during stereo matching and further improves the precision of the estimated disparities. Finally, the subpixel resolution is achieved by conducting a parabola interpolation and the subpixel disparity map is used to reconstruct the 3-D road surface. The proposed algorithm is implemented on an NVIDIA GTX 1080 GPU for the real-time purpose. The experimental results illustrate that the reconstruction accuracy is around 3 mm.

arXiv:1807.07132공식 코드 · Papers with Code

Newton-ADMM: A Distributed GPU-Accelerated Optimizer for Multiclass Classification Problems

First-order optimization methods, such as stochastic gradient descent (SGD) and its variants, are widely used in machine learning applications due to their simplicity and low per-iteration costs. However, they often require larger numbers of iterations, with associated communication costs in distributed environments. In contrast, Newton-type methods, while having higher per-iteration costs, typically require a significantly smaller number of iterations, which directly translates to reduced communication costs. In this paper, we present a novel distributed optimizer for classification problems, which integrates a GPU-accelerated Newton-type solver with the global consensus formulation of Alternating Direction of Method Multipliers (ADMM). By leveraging the communication efficiency of ADMM, GPU-accelerated inexact-Newton solver, and an effective spectral penalty parameter selection strategy, we show that our proposed method (i) yields better generalization performance on several classification problems; (ii) significantly outperforms state-of-the-art methods in distributed time to solution; and (iii) offers better scaling on large distributed platforms.

arXiv:1807.06555

Training Recurrent Neural Networks against Noisy Computations during Inference

We explore the robustness of recurrent neural networks when the computations within the network are noisy. One of the motivations for looking into this problem is to reduce the high power cost of conventional computing of neural network operations through the use of analog neuromorphic circuits. Traditional GPU/CPU-centered deep learning architectures exhibit bottlenecks in power-restricted applications, such as speech recognition in embedded systems. The use of specialized neuromorphic circuits, where analog signals passed through memory-cell arrays are sensed to accomplish matrix-vector multiplications, promises large power savings and speed gains but brings with it the problems of limited precision of computations and unavoidable analog noise. In this paper we propose a method, called {\em Deep Noise Injection training}, to train RNNs to obtain a set of weights/biases that is much more robust against noisy computation during inference. We explore several RNN architectures, such as vanilla RNN and long-short-term memories (LSTM), and show that after convergence of Deep Noise Injection training the set of trained weights/biases has more consistent performance over a wide range of noise powers entering the network during inference. Surprisingly, we find that Deep Noise Injection training improves overall performance of some networks even for numerically accurate inference.

arXiv:1807.06288공식 코드 · Papers with Code

PointSeg: Real-Time Semantic Segmentation Based on 3D LiDAR Point Cloud

In this paper, we propose PointSeg, a real-time end-to-end semantic segmentation method for road-objects based on spherical images. We take the spherical image, which is transformed from the 3D LiDAR point clouds, as input of the convolutional neural networks (CNNs) to predict the point-wise semantic map. To make PointSeg applicable on a mobile system, we build the model based on the light-weight network, SqueezeNet, with several improvements. It maintains a good balance between memory cost and prediction performance. Our model is trained on spherical images and label masks projected from the KITTI 3D object detection dataset. Experiments show that PointSeg can achieve competitive accuracy with 90fps on a single GPU 1080ti. which makes it quite compatible for autonomous driving applications.

arXiv:1807.05948

Evolving Differentiable Gene Regulatory Networks

Over the past twenty years, artificial Gene Regulatory Networks (GRNs) have shown their capacity to solve real-world problems in various domains such as agent control, signal processing and artificial life experiments. They have also benefited from new evolutionary approaches and improvements to dynamic which have increased their optimization efficiency. In this paper, we present an additional step toward their usability in machine learning applications. We detail an GPU-based implementation of differentiable GRNs, allowing for local optimization of GRN architectures with stochastic gradient descent (SGD). Using a standard machine learning dataset, we evaluate the ways in which evolution and SGD can be combined to further GRN optimization. We compare these approaches with neural network models trained by SGD and with support vector machines.

arXiv:1807.03954

Knowledge Extracted from Recurrent Deep Belief Network for Real Time Deterministic Control

Recently, the market on deep learning including not only software but also hardware is developing rapidly. Big data is collected through IoT devices and the industry world will analyze them to improve their manufacturing process. Deep Learning has the hierarchical network architecture to represent the complicated features of input patterns. Although deep learning can show the high capability of classification, prediction, and so on, the implementation on GPU devices are required. We may meet the trade-off between the higher precision by deep learning and the higher cost with GPU devices. We can success the knowledge extraction from the trained deep learning with high classification capability. The knowledge that can realize faster inference of pre-trained deep network is extracted as IF-THEN rules from the network signal flow given input data. Some experiment results with benchmark tests for time series data sets showed the effectiveness of our proposed method related to the computational speed.

arXiv:1807.03425

A GPU-Oriented Algorithm Design for Secant-Based Dimensionality Reduction

Dimensionality-reduction techniques are a fundamental tool for extracting useful information from high-dimensional data sets. Because secant sets encode manifold geometry, they are a useful tool for designing meaningful data-reduction algorithms. In one such approach, the goal is to construct a projection that maximally avoids secant directions and hence ensures that distinct data points are not mapped too close together in the reduced space. This type of algorithm is based on a mathematical framework inspired by the constructive proof of Whitney's embedding theorem from differential topology. Computing all (unit) secants for a set of points is by nature computationally expensive, thus opening the door for exploitation of GPU architecture for achieving fast versions of these algorithms. We present a polynomial-time data-reduction algorithm that produces a meaningful low-dimensional representation of a data set by iteratively constructing improved projections within the framework described above. Key to our algorithm design and implementation is the use of GPUs which, among other things, minimizes the computational time required for the calculation of all secant lines. One goal of this report is to share ideas with GPU experts and to discuss a class of mathematical algorithms that may be of interest to the broader GPU community.

arXiv:1807.02752

Real-time stereo vision-based lane detection system

The detection of multiple curved lane markings on a non-flat road surface is still a challenging task for automotive applications. To make an improvement, the depth information can be used to greatly enhance the robustness of the lane detection systems. The proposed system in this paper is developed from our previous work where the dense vanishing point Vp is estimated globally to assist the detection of multiple curved lane markings. However, the outliers in the optimal solution may severely affect the accuracy of the least squares fitting when estimating Vp. Therefore, in this paper we use Random Sample Consensus to update the inliers and outliers iteratively until the fraction of the number of inliers versus the total number exceeds our pre-set threshold. This significantly helps the system to overcome some suddenly changing conditions. Furthermore, we propose a novel lane position validation approach which provides a piecewise weight based on Vp and the gradient to reduce the gradient magnitude of the non-lane candidates. Then, we compute the energy of each possible solution and select all satisfying lane positions for visualisation. The proposed system is implemented on a heterogeneous system which consists of an Intel Core i7-4720HQ CPU and a NVIDIA GTX 970M GPU. A processing speed of 143 fps has been achieved, which is over 38 times faster than our previous work. Also, in order to evaluate the detection precision, we tested 2495 frames with 5361 lanes from the KITTI database (1637 lanes more than our previous experiment). It is shown that the overall successful detection rate is improved from 98.7% to 99.5%.

arXiv:1807.02587

Fast and Accurate Point Cloud Registration using Trees of Gaussian Mixtures

Point cloud registration sits at the core of many important and challenging 3D perception problems including autonomous navigation, SLAM, object/scene recognition, and augmented reality. In this paper, we present a new registration algorithm that is able to achieve state-of-the-art speed and accuracy through its use of a hierarchical Gaussian Mixture Model (GMM) representation. Our method constructs a top-down multi-scale representation of point cloud data by recursively running many small-scale data likelihood segmentations in parallel on a GPU. We leverage the resulting representation using a novel PCA-based optimization criterion that adaptively finds the best scale to perform data association between spatial subsets of point cloud data. Compared to previous Iterative Closest Point and GMM-based techniques, our tree-based point association algorithm performs data association in logarithmic-time while dynamically adjusting the level of detail to best match the complexity and spatial distribution characteristics of local scene geometry. In addition, unlike other GMM methods that restrict covariances to be isotropic, our new PCA-based optimization criterion well-approximates the true MLE solution even when fully anisotropic Gaussian covariances are used. Efficient data association, multi-scale adaptability, and a robust MLE approximation produce an algorithm that is up to an order of magnitude both faster and more accurate than current state-of-the-art on a wide variety of 3D datasets captured from LiDAR to structured light.

arXiv:1807.02894

Automatic Classification of Defective Photovoltaic Module Cells in Electroluminescence Images

Electroluminescence (EL) imaging is a useful modality for the inspection of photovoltaic (PV) modules. EL images provide high spatial resolution, which makes it possible to detect even finest defects on the surface of PV modules. However, the analysis of EL images is typically a manual process that is expensive, time-consuming, and requires expert knowledge of many different types of defects. In this work, we investigate two approaches for automatic detection of such defects in a single image of a PV cell. The approaches differ in their hardware requirements, which are dictated by their respective application scenarios. The more hardware-efficient approach is based on hand-crafted features that are classified in a Support Vector Machine (SVM). To obtain a strong performance, we investigate and compare various processing variants. The more hardware-demanding approach uses an end-to-end deep Convolutional Neural Network (CNN) that runs on a Graphics Processing Unit (GPU). Both approaches are trained on 1,968 cells extracted from high resolution EL intensity images of mono- and polycrystalline PV modules. The CNN is more accurate, and reaches an average accuracy of 88.42%. The SVM achieves a slightly lower average accuracy of 82.44%, but can run on arbitrary hardware. Both automated approaches make continuous, highly accurate monitoring of PV cells feasible.

arXiv:1807.02187

Encoding Motion Primitives for Autonomous Vehicles using Virtual Velocity Constraints and Neural Network Scheduling

Within the context of trajectory planning for autonomous vehicles this paper proposes methods for efficient encoding of motion primitives in neural networks on top of model-based and gradient-free reinforcement learning. It is distinguished between 5 core aspects: system model, network architecture, training algorithm, training tasks selection and hardware/software implementation. For the system model, a kinematic (3-states-2-controls) and a dynamic (16-states-2-controls) vehicle model are compared. For the network architecture, 3 feedforward structures are compared including weighted skip connections. For the training algorithm, virtual velocity constraints and network scheduling are proposed. For the training tasks, different feature vector selections are discussed. For the implementation, aspects of gradient-free learning using 1 GPU and the handling of perturbation noise therefore are discussed. The effects of proposed methods are illustrated in experiments encoding up to 14625 motion primitives. The capabilities of tiny neural networks with as few as 10 scalar parameters when scheduled on vehicle velocity are emphasized.

arXiv:1807.02004

Calamari - A High-Performance Tensorflow-based Deep Learning Package for Optical Character Recognition

Optical Character Recognition (OCR) on contemporary and historical data is still in the focus of many researchers. Especially historical prints require book specific trained OCR models to achieve applicable results (Springmann and L\"udeling, 2016, Reul et al., 2017a). To reduce the human effort for manually annotating ground truth (GT) various techniques such as voting and pretraining have shown to be very efficient (Reul et al., 2018a, Reul et al., 2018b). Calamari is a new open source OCR line recognition software that both uses state-of-the art Deep Neural Networks (DNNs) implemented in Tensorflow and giving native support for techniques such as pretraining and voting. The customizable network architectures constructed of Convolutional Neural Networks (CNNS) and Long-ShortTerm-Memory (LSTM) layers are trained by the so-called Connectionist Temporal Classification (CTC) algorithm of Graves et al. (2006). Optional usage of a GPU drastically reduces the computation times for both training and prediction. We use two different datasets to compare the performance of Calamari to OCRopy, OCRopus3, and Tesseract 4. Calamari reaches a Character Error Rate (CER) of 0.11% on the UW3 dataset written in modern English and 0.18% on the DTA19 dataset written in German Fraktur, which considerably outperforms the results of the existing softwares.

arXiv:1807.02044공식 코드 · Papers with Code

Real-Time Subpixel Fast Bilateral Stereo

Stereo vision technique has been widely used in robotic systems to acquire 3-D information. In recent years, many researchers have applied bilateral filtering in stereo vision to adaptively aggregate the matching costs. This has greatly improved the accuracy of the estimated disparity maps. However, the process of filtering the whole cost volume is very time consuming and therefore the researchers have to resort to some powerful hardware for the real-time purpose. This paper presents the implementation of fast bilateral stereo on a state-of-the-art GPU. By highly exploiting the parallel computing architecture of the GPU, the fast bilateral stereo performs in real time when processing the Middlebury stereo datasets.

arXiv:1806.11248공식 코드 · Papers with Code

XGBoost: Scalable GPU Accelerated Learning

We describe the multi-GPU gradient boosting algorithm implemented in the XGBoost library (https://github.com/dmlc/xgboost). Our algorithm allows fast, scalable training on multi-GPU systems with all of the features of the XGBoost library. We employ data compression techniques to minimise the usage of scarce GPU memory while still allowing highly efficient implementation. Using our algorithm we show that it is possible to process 115 million training instances in under three minutes on a publicly available cloud computing instance. The algorithm is implemented using end-to-end GPU parallelism, with prediction, gradient calculation, feature quantisation, decision tree construction and evaluation phases all computed on device.

arXiv:1806.10779

Differentiable Learning-to-Normalize via Switchable Normalization

We address a learning-to-normalize problem by proposing Switchable Normalization (SN), which learns to select different normalizers for different normalization layers of a deep neural network. SN employs three distinct scopes to compute statistics (means and variances) including a channel, a layer, and a minibatch. SN switches between them by learning their importance weights in an end-to-end manner. It has several good properties. First, it adapts to various network architectures and tasks (see Fig.1). Second, it is robust to a wide range of batch sizes, maintaining high performance even when small minibatch is presented (e.g. 2 images/GPU). Third, SN does not have sensitive hyper-parameter, unlike group normalization that searches the number of groups as a hyper-parameter. Without bells and whistles, SN outperforms its counterparts on various challenging benchmarks, such as ImageNet, COCO, CityScapes, ADE20K, and Kinetics. Analyses of SN are also presented. We hope SN will help ease the usage and understand the normalization techniques in deep learning. The code of SN has been made available in https://github.com/switchablenorms/.

arXiv:1806.10342공식 코드 · Papers with Code

3D RoI-aware U-Net for Accurate and Efficient Colorectal Tumor Segmentation

Segmentation of colorectal cancerous regions from 3D Magnetic Resonance (MR) images is a crucial procedure for radiotherapy which conventionally requires accurate delineation of tumour boundaries at an expense of labor, time and reproducibility. While deep learning based methods serve good baselines in 3D image segmentation tasks, small applicable patch size limits effective receptive field and degrades segmentation performance. In addition, Regions of interest (RoIs) localization from large whole volume 3D images serves as a preceding operation that brings about multiple benefits in terms of speed, target completeness, reduction of false positives. Distinct from sliding window or non-joint localization-segmentation based models, we propose a novel multitask framework referred to as 3D RoI-aware U-Net (3D RU-Net), for RoI localization and in-region segmentation where the two tasks share one backbone encoder network. With the region proposals from the encoder, we crop multi-level RoI in-region features from the encoder to form a GPU memory-efficient decoder for detailpreserving segmentation and therefore enlarged applicable volume size and effective receptive field. To effectively train the model, we designed a Dice formulated loss function for the global-to-local multi-task learning procedure. Based on the efficiency gains, we went on to ensemble models with different receptive fields to achieve even higher performance costing minor extra computational expensiveness. Extensive experiments were conducted on 64 cancerous cases with a four-fold cross-validation, and the results showed significant superiority in terms of accuracy and efficiency over conventional frameworks. In conclusion, the proposed method has a huge potential for extension to other 3D object segmentation tasks from medical images due to its inherent generalizability. The code for the proposed method is publicly available.

arXiv:1806.10282

Auto-Keras: An Efficient Neural Architecture Search System

Neural architecture search (NAS) has been proposed to automatically tune deep neural networks, but existing search algorithms, e.g., NASNet, PNAS, usually suffer from expensive computational cost. Network morphism, which keeps the functionality of a neural network while changing its neural architecture, could be helpful for NAS by enabling more efficient training during the search. In this paper, we propose a novel framework enabling Bayesian optimization to guide the network morphism for efficient neural architecture search. The framework develops a neural network kernel and a tree-structured acquisition function optimization algorithm to efficiently explores the search space. Intensive experiments on real-world benchmark datasets have been done to demonstrate the superior performance of the developed framework over the state-of-the-art methods. Moreover, we build an open-source AutoML system based on our method, namely Auto-Keras. The system runs in parallel on CPU and GPU, with an adaptive search strategy for different GPU memory limits.

arXiv:1806.10170

Detection of Alzheimers Disease from MRI using Convolutional Neural Network with Tensorflow

Nowadays, due to tremendous improvements in high performance computing, it has become easier to train Neural Networks. We intend to take advantage of this situation and apply this technology in solving real world problems. There was a need for automatic diagnosis certain diseases from medical images that could help a doctor and radiologist for further action towards treating the illness. We chose Alzheimer disease for this purpose. Alzheimer disease is the leading cause of dementia and memory loss. Alzheimer disease, it is caused by atrophy of the certain brain regions and by brain cell death. MRI scans reveal this information but atrophy regions are different for different people which makes the diagnosis a little trickier and often gets miss-diagnosed by doctors and radiologists. The Dataset used for this project is provided by OASIS, which contains over 400 subjects 100 of which having mild to severe dementia and is supplemented by MMSE and CDR standards of diagnosis in the same context. Enter CNN, Convolutional Neural Networks are a hybrid of Kernel Convolutions and Neural Networks. Kernel Convolutions is a technique that uses filters to recognize and segment images based on features. Neural Networks consist of neurons which are loosely based on human brains neuron which represents a single classifier and interconnected by weights, have different biases and are activated by some activation functions. By using Convolutional Neural Networks, the problem can be solved with minimal error rate. The technologies we intend to use are libraries like CUDA CuDNN for making use of GPU and its multiple cores-parallel computing to train models while giving us high performance.