Subsampling Operations Research Articles

Hierarchical, rotation-equivariant neural networks to select structural models of protein complexes.

Predicting the structure of multi-protein complexes is a grand challenge in biochemistry, with major implications for basic science and drug discovery. Computational structure prediction methods generally leverage predefined structural features to distinguish accurate structural models from less accurate ones. This raises the question of whether it is possible to learn characteristics of accurate models directly from atomic coordinates of protein complexes, with no prior assumptions. Here we introduce a machine learning method that learns directly from the 3D positions of all atoms to identify accurate models of protein complexes, without using any precomputed physics-inspired or statistical terms. Our neural network architecture combines multiple ingredients that together enable end-to-end learning from molecular structures containing tens of thousands of atoms: a point-based representation of atoms, equivariance with respect to rotation and translation, local convolutions, and hierarchical subsampling operations. When used in combination with previously developed scoring functions, our network substantially improves the identification of accurate structural models among a large set of possible models. Our network can also be used to predict the accuracy of a given structural model in absolute terms. The architecture we present is readily applicable to other tasks involving learning on 3D structures of large atomic systems.

Salient object detection via light-weight multi-path cascaded networks

Recently, deep learning-based saliency detection has achieved fantastic performance over conventional works. However, repeated subsampling operations like pooling or convolution striding in deep CNNs lead to a significant decrease in the initial image resolution. It is difficult for a network to learn saliency at the boundaries of salient regions. In this paper, a Light-Weight Multi-Path Cascaded Network (LMCN) is proposed for image saliency detection – an encoder-decoder architecture that explicitly exploits all the information available along the down-sampling process to enable full-resolution prediction using long-range residual connections. The Res2Net is adopted as a backbone network to better extract multi-level and multi-scale feature maps by constructing hierarchical residual-like connections within one single residual block. Also, the network is lightweighted by the bottleneck structure and depth-wise convolution, which ultimately reduces the number of parameters by more than 50% while maintaining the same performance. Comprehensive experiments are carried out and new state-of-the-art results are set in eight public datasets. Experiment results show that the proposed approach substantially improves previous state-of-the-art performances.

Learning residual refinement network with semantic context representation for real-time saliency object detection

Salient object detection (SOD) aims to precisely segment out the most attractive areas in a single image. With the rapid development of deep learning, much effort has been paid to learn an effective representation for SOD from bottom-up or top-down pathways. However, they fail to precisely separate out the whole salient object with fine boundaries due to the repeated subsampling operations such as pooling and striding leading to the loss of fine structures and spatial details. To address these issues, in this paper, we propose a residual refinement network with semantic context features for SOD. First, we design an encoder-decoder structure with side-connections to capture the sharper object boundaries, which can not only gradually recover the spatial details in each feature map from top to down, but also enhance the features at all scales with high-level semantic context information. The semantic context enhanced features are further strengthen by using a set of atrous convolutional filters with multiple atrous rates to encode multi-scale context information. Finally, using the side-output features as input, we develop a recurrent residual module to gradually learn to recover the missing boundary details in the previous coarsely predicted saliency map in a coarse-to-fine manner. Extensive evaluations on six popular SOD benchmark datasets demonstrate leading performance of the proposed approach compared with state-of-the-art methods. Especially, our approach runs in real-time at a speed of 29 fps.

A Multistage Refinement Network for Salient Object Detection.

Deep convolutional neural networks (CNNs) have been successfully applied to a wide variety of problems in computer vision, including salient object detection. To accurately detect and segment salient objects, it is necessary to extract and combine high-level semantic features with low-level fine details simultaneously. This is challenging for CNNs because repeated subsampling operations such as pooling and convolution lead to a significant decrease in the feature resolution, which results in the loss of spatial details and finer structures. Therefore, we propose augmenting feedforward neural networks by using the multistage refinement mechanism. In the first stage, a master net is built to generate a coarse prediction map in which most detailed structures are missing. In the following stages, the refinement net with layerwise recurrent connections to the master net is equipped to progressively combine local context information across stages to refine the preceding saliency maps in a stagewise manner. Furthermore, the pyramid pooling module and channel attention module are applied to aggregate different-region-based global contexts. Extensive evaluations over six benchmark datasets show that the proposed method performs favorably against the state-of-the-art approaches.

A minimum entropy deconvolution-enhanced convolutional neural networks for fault diagnosis of axial piston pumps

Condition monitoring of piston pumps has great significance to ensure the reliability and security of hydraulic systems. However, the complex working conditions of the integrated electromechanical systems make the fault mechanism unclear which is difficult for fault diagnosis by feature matching techniques. In this paper, a novel minimum entropy deconvolution (MED)-based convolutional neural network (CNN) is presented to classify faults in axial piston pumps. Firstly, the collected raw signals are preprocessed using the MED technique. Then, the filtered signals are used to construct training samples and testing samples. Finally, the constructed samples are fed into the CNN to classify the multi-faults of axial piston pumps. With the convolution and subsampling operations, the present model can automatically obtain data features via iterative learning processes, which is suitable for the unknown fault mechanism problems. The learned features are visualized by t-distributed stochastic neighbor embedding technique. A benchmark data simulation of mechanical transmission systems and an experimental data investigation of an axial piston pump are performed to manifest the superiority of the present method by comparing with the traditional CNN.

VoxSegNet: Volumetric CNNs for Semantic Part Segmentation of 3D Shapes.

Volumetric representation has been widely used for 3D deep learning in shape analysis due to its generalization ability and regular data format. However, for fine-grained tasks like part segmentation, volumetric data has not been widely adopted compared to other representations. Aiming at delivering an effective volumetric method for 3D shape part segmentation, this paper proposes a novel volumetric convolutional neural network. Our method can extract discriminative features encoding detailed information from voxelized 3D data under limited resolution. To this purpose, a spatial dense extraction (SDE) module is designed to preserve spatial resolution during feature extraction procedure, alleviating the loss of details caused by sub-sampling operations such as max pooling. An attention feature aggregation (AFA) module is also introduced to adaptively select informative features from different abstraction levels, leading to segmentation with both semantic consistency and high accuracy of details. Experimental results demonstrate that promising results can be achieved by using volumetric data, with part segmentation accuracy comparable or superior to state-of-the-art non-volumetric methods.

Flattenet: A Simple and Versatile Framework for Dense Pixelwise Prediction

In this paper, we focus on devising a versatile framework for dense pixelwise prediction whose goal is to assign a discrete or continuous label to each pixel for an image. It is well-known that the reduced feature resolution due to repeated subsampling operations poses a serious challenge to Fully Convolutional Network (FCN) based models. In contrast to the commonly-used strategies, such as dilated convolution and encoder-decoder structure, we introduce the Flattening Module to produce high-resolution predictions without either removing any subsampling operations or building a complicated decoder module. In addition, the Flattening Module is lightweight and can be easily combined with any existing FCNs, allowing the model builder to trade off among model size, computational cost and accuracy by simply choosing different backbone networks. We empirically demonstrate the effectiveness of the proposed Flattening Module through competitive results in human pose estimation on MPII, semantic segmentation on PASCAL-Context and object detection on PASCAL VOC. We hope that the proposed approach can serve as a simple and strong alternative of current dominant dense pixelwise prediction frameworks.

RefineNet: Multi-Path Refinement Networks for Dense Prediction.

Recently, very deep convolutional neural networks (CNNs) have shown outstanding performance in object recognition and have also been the first choice for dense prediction problems such as semantic segmentation and depth estimation. However, repeated subsampling operations like pooling or convolution striding in deep CNNs lead to a significant decrease in the initial image resolution. Here, we present RefineNet, a generic multi-path refinement network that explicitly exploits all the information available along the down-sampling process to enable high-resolution prediction using long-range residual connections. In this way, the deeper layers that capture high-level semantic features can be directly refined using fine-grained features from earlier convolutions. The individual components of RefineNet employ residual connections following the identity mapping mindset, which allows for effective end-to-end training. Further, we introduce chained residual pooling, which captures rich background context in an efficient manner. We carry out comprehensive experiments on semantic segmentation which is a dense classification problem and achieve good performance on seven public datasets. We further apply our method for depth estimation and demonstrate the effectiveness of our method on dense regression problems.

Refinet: A Deep Segmentation Assisted Refinement Network for Salient Object Detection

Compared to conventional saliency detection by handcrafted features, deep convolutional neural networks (CNNs) recently have been successfully applied to saliency detection field with superior performance on locating salient objects. However, due to repeated sub-sampling operations inside CNNs such as pooling and convolution, many CNN-based saliency models fail to maintain fine-grained spatial details and boundary structures of objects. To remedy this issue, this paper proposes a novel end-to-end deep learning-based refinement model named Refinet , which is based on fully convolutional network augmented with segmentation hypotheses. Intermediate saliency maps that are edge-aware are computed from segmentation-based pooling and then feed to a two-tier fully convolutional network for effective fusion and refinement, leading to more precise object details and boundaries. In addition, the resolution of feature maps in the proposed Refinet is carefully designed to guarantee sufficient boundary clarity of the refined saliency output. Compared to widely employed dense conditional random field, Refinet is able to enhance coarse saliency maps generated by existing models with more accurate spatial details, and its effectiveness is demonstrated by experimental results on seven benchmark datasets.

Quantization Noise Analysis of Time-to-Digital-Converter-Based All-Digital Phase-Locked Loop and Frequency Discriminators

All-digital phase-locked loops (ADPLLs) based on the time-to-digital converter (TDC) and the frequency discriminator (FD) are modeled and analyzed in terms of quantization effects. Using linear models with quantization noise sources, theoretical analysis and simulation are carried out to obtain the output phase noise of each building block of the TDC-based ADPLL. It is newly derived that the TDC noise component caused by the delta-sigma modulator (DSM) and the finite resolution of the digitally controlled oscillator is not white. Namely, the in-band phase noise caused by the DSM-induced TDC is not white, which is due to the integrate-and-dump and subsampling operations of the TDC. This can give some insight into the design of low-noise ADPLLs. Some structures of delta-sigma FDs, which can serve as an alternative to the TDC, are also newly analyzed in terms of quantization noise, using the derived linear noise model.

Fast and Robust Face Detection on a Parallel Optimized Architecture Implemented on FPGA

In this paper, we present a parallel architecture for fast and robust face detection implemented on FPGA hardware. We propose the first implementation that meets both real-time requirements in an embedded context and face detection robustness within complex backgrounds. The chosen face detection method is the Convolutional Face Finder (CFF) algorithm, which consists of a pipeline of convolution and subsampling operations, followed by a multilayer perceptron. We present the design methodology of our face detection processor element (PE). This methodology was followed in order to optimize our implementation in terms of memory usage and parallelization efficiency. We then built a parallel architecture composed of a PE ring and an FIFO memory, resulting in a scalable system capable of processing images of different sizes. A ring of 25 PEs running at 80 MHz is able to process 127 QVGA images per second and performing real-time face detection on VGA images (35 images per second).