PyTorchKR Papers

최신 논문

총 2,147편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2008.05402

DAWN: Vehicle Detection in Adverse Weather Nature Dataset

Recently, self-driving vehicles have been introduced with several automated features including lane-keep assistance, queuing assistance in traffic-jam, parking assistance and crash avoidance. These self-driving vehicles and intelligent visual traffic surveillance systems mainly depend on cameras and sensors fusion systems. Adverse weather conditions such as heavy fog, rain, snow, and sandstorms are considered dangerous restrictions of the functionality of cameras impacting seriously the performance of adopted computer vision algorithms for scene understanding (i.e., vehicle detection, tracking, and recognition in traffic scenes). For example, reflection coming from rain flow and ice over roads could cause massive detection errors which will affect the performance of intelligent visual traffic systems. Additionally, scene understanding and vehicle detection algorithms are mostly evaluated using datasets contain certain types of synthetic images plus a few real-world images. Thus, it is uncertain how these algorithms would perform on unclear images acquired in the wild and how the progress of these algorithms is standardized in the field. To this end, we present a new dataset (benchmark) consisting of real-world images collected under various adverse weather conditions called DAWN. This dataset emphasizes a diverse traffic environment (urban, highway and freeway) as well as a rich variety of traffic flow. The DAWN dataset comprises a collection of 1000 images from real-traffic environments, which are divided into four sets of weather conditions: fog, snow, rain and sandstorms. The dataset is annotated with object bounding boxes for autonomous driving and video surveillance scenarios. This data helps interpreting effects caused by the adverse weather conditions on the performance of vehicle detection systems.

arXiv:2008.05309

Factor Graph based 3D Multi-Object Tracking in Point Clouds

Accurate and reliable tracking of multiple moving objects in 3D space is an essential component of urban scene understanding. This is a challenging task because it requires the assignment of detections in the current frame to the predicted objects from the previous one. Existing filter-based approaches tend to struggle if this initial assignment is not correct, which can happen easily. We propose a novel optimization-based approach that does not rely on explicit and fixed assignments. Instead, we represent the result of an off-the-shelf 3D object detector as Gaussian mixture model, which is incorporated in a factor graph framework. This gives us the flexibility to assign all detections to all objects simultaneously. As a result, the assignment problem is solved implicitly and jointly with the 3D spatial multi-object state estimation using non-linear least squares optimization. Despite its simplicity, the proposed algorithm achieves robust and reliable tracking results and can be applied for offline as well as online tracking. We demonstrate its performance on the real world KITTI tracking dataset and achieve better results than many state-of-the-art algorithms. Especially the consistency of the estimated tracks is superior offline as well as online.

arXiv:2008.02986

Global Context Aware Convolutions for 3D Point Cloud Understanding

Recent advances in deep learning for 3D point clouds have shown great promises in scene understanding tasks thanks to the introduction of convolution operators to consume 3D point clouds directly in a neural network. Point cloud data, however, could have arbitrary rotations, especially those acquired from 3D scanning. Recent works show that it is possible to design point cloud convolutions with rotation invariance property, but such methods generally do not perform as well as translation-invariant only convolution. We found that a key reason is that compared to point coordinates, rotation-invariant features consumed by point cloud convolution are not as distinctive. To address this problem, we propose a novel convolution operator that enhances feature distinction by integrating global context information from the input point cloud to the convolution. To this end, a globally weighted local reference frame is constructed in each point neighborhood in which the local point set is decomposed into bins. Anchor points are generated in each bin to represent global shape features. A convolution can then be performed to transform the points and anchor features into final rotation-invariant features. We conduct several experiments on point cloud classification, part segmentation, shape retrieval, and normals estimation to evaluate our convolution, which achieves state-of-the-art accuracy under challenging rotations.

arXiv:2008.02918공식 코드 · Papers with Code

Polysemy Deciphering Network for Robust Human-Object Interaction Detection

Human-Object Interaction (HOI) detection is important to human-centric scene understanding tasks. Existing works tend to assume that the same verb has similar visual characteristics in different HOI categories, an approach that ignores the diverse semantic meanings of the verb. To address this issue, in this paper, we propose a novel Polysemy Deciphering Network (PD-Net) that decodes the visual polysemy of verbs for HOI detection in three distinct ways. First, we refine features for HOI detection to be polysemyaware through the use of two novel modules: namely, Language Prior-guided Channel Attention (LPCA) and Language Prior-based Feature Augmentation (LPFA). LPCA highlights important elements in human and object appearance features for each HOI category to be identified; moreover, LPFA augments human pose and spatial features for HOI detection using language priors, enabling the verb classifiers to receive language hints that reduce intra-class variation for the same verb. Second, we introduce a novel Polysemy-Aware Modal Fusion module (PAMF), which guides PD-Net to make decisions based on feature types deemed more important according to the language priors. Third, we propose to relieve the verb polysemy problem through sharing verb classifiers for semantically similar HOI categories. Furthermore, to expedite research on the verb polysemy problem, we build a new benchmark dataset named HOI-VerbPolysemy (HOIVP), which includes common verbs (predicates) that have diverse semantic meanings in the real world. Finally, through deciphering the visual polysemy of verbs, our approach is demonstrated to outperform state-of-the-art methods by significant margins on the HICO-DET, V-COCO, and HOI-VP databases. Code and data in this paper are available at https://github.com/MuchHair/PD-Net.

arXiv:2008.02042공식 코드 · Papers with Code

Pose-based Modular Network for Human-Object Interaction Detection

Human-object interaction(HOI) detection is a critical task in scene understanding. The goal is to infer the triplet <subject, predicate, object> in a scene. In this work, we note that the human pose itself as well as the relative spatial information of the human pose with respect to the target object can provide informative cues for HOI detection. We contribute a Pose-based Modular Network (PMN) which explores the absolute pose features and relative spatial pose features to improve HOI detection and is fully compatible with existing networks. Our module consists of a branch that first processes the relative spatial pose features of each joint independently. Another branch updates the absolute pose features via fully connected graph structures. The processed pose features are then fed into an action classifier. To evaluate our proposed method, we combine the module with the state-of-the-art model named VS-GATs and obtain significant improvement on two public benchmarks: V-COCO and HICO-DET, which shows its efficacy and flexibility. Code is available at \url{https://github.com/birlrobotics/PMN}.

arXiv:2007.13970공식 코드 · Papers with Code

Weakly Supervised 3D Object Detection from Point Clouds

A crucial task in scene understanding is 3D object detection, which aims to detect and localize the 3D bounding boxes of objects belonging to specific classes. Existing 3D object detectors heavily rely on annotated 3D bounding boxes during training, while these annotations could be expensive to obtain and only accessible in limited scenarios. Weakly supervised learning is a promising approach to reducing the annotation requirement, but existing weakly supervised object detectors are mostly for 2D detection rather than 3D. In this work, we propose VS3D, a framework for weakly supervised 3D object detection from point clouds without using any ground truth 3D bounding box for training. First, we introduce an unsupervised 3D proposal module that generates object proposals by leveraging normalized point cloud densities. Second, we present a cross-modal knowledge distillation strategy, where a convolutional neural network learns to predict the final results from the 3D object proposals by querying a teacher network pretrained on image datasets. Comprehensive experiments on the challenging KITTI dataset demonstrate the superior performance of our VS3D in diverse evaluation settings. The source code and pretrained models are publicly available at https://github.com/Zengyi-Qin/Weakly-Supervised-3D-Object-Detection.

arXiv:2007.13138

Virtual Multi-view Fusion for 3D Semantic Segmentation

Semantic segmentation of 3D meshes is an important problem for 3D scene understanding. In this paper we revisit the classic multiview representation of 3D meshes and study several techniques that make them effective for 3D semantic segmentation of meshes. Given a 3D mesh reconstructed from RGBD sensors, our method effectively chooses different virtual views of the 3D mesh and renders multiple 2D channels for training an effective 2D semantic segmentation model. Features from multiple per view predictions are finally fused on 3D mesh vertices to predict mesh semantic segmentation labels. Using the large scale indoor 3D semantic segmentation benchmark of ScanNet, we show that our virtual views enable more effective training of 2D semantic segmentation networks than previous multiview approaches. When the 2D per pixel predictions are aggregated on 3D surfaces, our virtual multiview fusion method is able to achieve significantly better 3D semantic segmentation results compared to all prior multiview approaches and competitive with recent 3D convolution approaches.

arXiv:2007.12868

OpenRooms: An End-to-End Open Framework for Photorealistic Indoor Scene Datasets

We propose a novel framework for creating large-scale photorealistic datasets of indoor scenes, with ground truth geometry, material, lighting and semantics. Our goal is to make the dataset creation process widely accessible, transforming scans into photorealistic datasets with high-quality ground truth for appearance, layout, semantic labels, high quality spatially-varying BRDF and complex lighting, including direct, indirect and visibility components. This enables important applications in inverse rendering, scene understanding and robotics. We show that deep networks trained on the proposed dataset achieve competitive performance for shape, material and lighting estimation on real images, enabling photorealistic augmented reality applications, such as object insertion and material editing. We also show our semantic labels may be used for segmentation and multi-task learning. Finally, we demonstrate that our framework may also be integrated with physics engines, to create virtual robotics environments with unique ground truth such as friction coefficients and correspondence to real scenes. The dataset and all the tools to create such datasets will be made publicly available.

arXiv:2007.12107

Few-Shot Object Detection and Viewpoint Estimation for Objects in the Wild

Detecting objects and estimating their viewpoints in images are key tasks of 3D scene understanding. Recent approaches have achieved excellent results on very large benchmarks for object detection and viewpoint estimation. However, performances are still lagging behind for novel object categories with few samples. In this paper, we tackle the problems of few-shot object detection and few-shot viewpoint estimation. We demonstrate on both tasks the benefits of guiding the network prediction with class-representative features extracted from data in different modalities: image patches for object detection, and aligned 3D models for viewpoint estimation. Despite its simplicity, our method outperforms state-of-the-art methods by a large margin on a range of datasets, including PASCAL and COCO for few-shot object detection, and Pascal3D+ and ObjectNet3D for few-shot viewpoint estimation. Furthermore, when the 3D model is not available, we introduce a simple category-agnostic viewpoint estimation method by exploiting geometrical similarities and consistent pose labelling across different classes. While it moderately reduces performance, this approach still obtains better results than previous methods in this setting. Last, for the first time, we tackle the combination of both few-shot tasks, on three challenging benchmarks for viewpoint estimation in the wild, ObjectNet3D, Pascal3D+ and Pix3D, showing very promising results.

arXiv:2007.10985공식 코드 · Papers with Code

PointContrast: Unsupervised Pre-training for 3D Point Cloud Understanding

Arguably one of the top success stories of deep learning is transfer learning. The finding that pre-training a network on a rich source set (eg., ImageNet) can help boost performance once fine-tuned on a usually much smaller target set, has been instrumental to many applications in language and vision. Yet, very little is known about its usefulness in 3D point cloud understanding. We see this as an opportunity considering the effort required for annotating data in 3D. In this work, we aim at facilitating research on 3D representation learning. Different from previous works, we focus on high-level scene understanding tasks. To this end, we select a suite of diverse datasets and tasks to measure the effect of unsupervised pre-training on a large source set of 3D scenes. Our findings are extremely encouraging: using a unified triplet of architecture, source dataset, and contrastive loss for pre-training, we achieve improvement over recent best results in segmentation and detection across 6 different benchmarks for indoor and outdoor, real and synthetic datasets -- demonstrating that the learned representation can generalize across domains. Furthermore, the improvement was similar to supervised pre-training, suggesting that future efforts should favor scaling data collection over more detailed annotation. We hope these findings will encourage more research on unsupervised pretext task design for 3D deep learning.

arXiv:2007.09107

Synthetic and Real Inputs for Tool Segmentation in Robotic Surgery

Semantic tool segmentation in surgical videos is important for surgical scene understanding and computer-assisted interventions as well as for the development of robotic automation. The problem is challenging because different illumination conditions, bleeding, smoke and occlusions can reduce algorithm robustness. At present labelled data for training deep learning models is still lacking for semantic surgical instrument segmentation and in this paper we show that it may be possible to use robot kinematic data coupled with laparoscopic images to alleviate the labelling problem. We propose a new deep learning based model for parallel processing of both laparoscopic and simulation images for robust segmentation of surgical tools. Due to the lack of laparoscopic frames annotated with both segmentation ground truth and kinematic information a new custom dataset was generated using the da Vinci Research Kit (dVRK) and is made available.

arXiv:2007.04954공식 코드 · Papers with Code

ThreeDWorld: A Platform for Interactive Multi-Modal Physical Simulation

We introduce ThreeDWorld (TDW), a platform for interactive multi-modal physical simulation. TDW enables simulation of high-fidelity sensory data and physical interactions between mobile agents and objects in rich 3D environments. Unique properties include: real-time near-photo-realistic image rendering; a library of objects and environments, and routines for their customization; generative procedures for efficiently building classes of new environments; high-fidelity audio rendering; realistic physical interactions for a variety of material types, including cloths, liquid, and deformable objects; customizable agents that embody AI agents; and support for human interactions with VR devices. TDW's API enables multiple agents to interact within a simulation and returns a range of sensor and physics data representing the state of the world. We present initial experiments enabled by TDW in emerging research directions in computer vision, machine learning, and cognitive science, including multi-modal physical scene understanding, physical dynamics predictions, multi-agent interactions, models that learn like a child, and attention studies in humans and neural networks.

arXiv:2007.03357공식 코드 · Papers with Code

Learning and Reasoning with the Graph Structure Representation in Robotic Surgery

Learning to infer graph representations and performing spatial reasoning in a complex surgical environment can play a vital role in surgical scene understanding in robotic surgery. For this purpose, we develop an approach to generate the scene graph and predict surgical interactions between instruments and surgical region of interest (ROI) during robot-assisted surgery. We design an attention link function and integrate with a graph parsing network to recognize the surgical interactions. To embed each node with corresponding neighbouring node features, we further incorporate SageConv into the network. The scene graph generation and active edge classification mostly depend on the embedding or feature extraction of node and edge features from complex image representation. Here, we empirically demonstrate the feature extraction methods by employing label smoothing weighted loss. Smoothing the hard label can avoid the over-confident prediction of the model and enhances the feature representation learned by the penultimate layer. To obtain the graph scene label, we annotate the bounding box and the instrument-ROI interactions on the robotic scene segmentation challenge 2018 dataset with an experienced clinical expert in robotic surgery and employ it to evaluate our propositions.

arXiv:2007.01151

JUMPS: Joints Upsampling Method for Pose Sequences

Human Pose Estimation is a low-level task useful forsurveillance, human action recognition, and scene understandingat large. It also offers promising perspectives for the animationof synthetic characters. For all these applications, and especiallythe latter, estimating the positions of many joints is desirablefor improved performance and realism. To this purpose, wepropose a novel method called JUMPS for increasing the numberof joints in 2D pose estimates and recovering occluded ormissing joints. We believe this is the first attempt to addressthe issue. We build on a deep generative model that combines aGenerative Adversarial Network (GAN) and an encoder. TheGAN learns the distribution of high-resolution human posesequences, the encoder maps the input low-resolution sequencesto its latent space. Inpainting is obtained by computing the latentrepresentation whose decoding by the GAN generator optimallymatches the joints locations at the input. Post-processing a 2Dpose sequence using our method provides a richer representationof the character motion. We show experimentally that thelocalization accuracy of the additional joints is on average onpar with the original pose estimates.

arXiv:2006.15618

Analogical Image Translation for Fog Generation

Image-to-image translation is to map images from a given \emph{style} to another given \emph{style}. While exceptionally successful, current methods assume the availability of training images in both source and target domains, which does not always hold in practice. Inspired by humans' reasoning capability of analogy, we propose analogical image translation (AIT). Given images of two styles in the source domain: $\mathcal{A}$ and $\mathcal{A}^\prime$, along with images $\mathcal{B}$ of the first style in the target domain, learn a model to translate $\mathcal{B}$ to $\mathcal{B}^\prime$ in the target domain, such that $\mathcal{A}:\mathcal{A}^\prime ::\mathcal{B}:\mathcal{B}^\prime$. AIT is especially useful for translation scenarios in which training data of one style is hard to obtain but training data of the same two styles in another domain is available. For instance, in the case from normal conditions to extreme, rare conditions, obtaining real training images for the latter case is challenging but obtaining synthetic data for both cases is relatively easy. In this work, we are interested in adding adverse weather effects, more specifically fog effects, to images taken in clear weather. To circumvent the challenge of collecting real foggy images, AIT learns with synthetic clear-weather images, synthetic foggy images and real clear-weather images to add fog effects onto real clear-weather images without seeing any real foggy images during training. AIT achieves this zero-shot image translation capability by coupling a supervised training scheme in the synthetic domain, a cycle consistency strategy in the real domain, an adversarial training scheme between the two domains, and a novel network design. Experiments show the effectiveness of our method for zero-short image translation and its benefit for downstream tasks such as semantic foggy scene understanding.

arXiv:2006.12567공식 코드 · Papers with Code

A Survey on Deep Learning for Localization and Mapping: Towards the Age of Spatial Machine Intelligence

Deep learning based localization and mapping has recently attracted significant attention. Instead of creating hand-designed algorithms through exploitation of physical models or geometric theories, deep learning based solutions provide an alternative to solve the problem in a data-driven way. Benefiting from ever-increasing volumes of data and computational power, these methods are fast evolving into a new area that offers accurate and robust systems to track motion and estimate scenes and their structure for real-world applications. In this work, we provide a comprehensive survey, and propose a new taxonomy for localization and mapping using deep learning. We also discuss the limitations of current models, and indicate possible future directions. A wide range of topics are covered, from learning odometry estimation, mapping, to global localization and simultaneous localization and mapping (SLAM). We revisit the problem of perceiving self-motion and scene understanding with on-board sensors, and show how to solve it by integrating these modules into a prospective spatial machine intelligence system (SMIS). It is our hope that this work can connect emerging works from robotics, computer vision and machine learning communities, and serve as a guide for future researchers to apply deep learning to tackle localization and mapping problems.

arXiv:2006.09623

Learning Visual Commonsense for Robust Scene Graph Generation

Scene graph generation models understand the scene through object and predicate recognition, but are prone to mistakes due to the challenges of perception in the wild. Perception errors often lead to nonsensical compositions in the output scene graph, which do not follow real-world rules and patterns, and can be corrected using commonsense knowledge. We propose the first method to acquire visual commonsense such as affordance and intuitive physics automatically from data, and use that to improve the robustness of scene understanding. To this end, we extend Transformer models to incorporate the structure of scene graphs, and train our Global-Local Attention Transformer on a scene graph corpus. Once trained, our model can be applied on any scene graph generation model and correct its obvious mistakes, resulting in more semantically plausible scene graphs. Through extensive experiments, we show our model learns commonsense better than any alternative, and improves the accuracy of state-of-the-art scene graph generation methods.

arXiv:2006.07744

Exploiting the ConvLSTM: Human Action Recognition using Raw Depth Video-Based Recurrent Neural Networks

As in many other different fields, deep learning has become the main approach in most computer vision applications, such as scene understanding, object recognition, computer-human interaction or human action recognition (HAR). Research efforts within HAR have mainly focused on how to efficiently extract and process both spatial and temporal dependencies of video sequences. In this paper, we propose and compare, two neural networks based on the convolutional long short-term memory unit, namely ConvLSTM, with differences in the architecture and the long-term learning strategy. The former uses a video-length adaptive input data generator (\emph{stateless}) whereas the latter explores the \emph{stateful} ability of general recurrent neural networks but applied in the particular case of HAR. This stateful property allows the model to accumulate discriminative patterns from previous frames without compromising computer memory. Experimental results on the large-scale NTU RGB+D dataset show that the proposed models achieve competitive recognition accuracies with lower computational cost compared with state-of-the-art methods and prove that, in the particular case of videos, the rarely-used stateful mode of recurrent neural networks significantly improves the accuracy obtained with the standard mode. The recognition accuracies obtained are 75.26\% (CS) and 75.45\% (CV) for the stateless model, with an average time consumption per video of 0.21 s, and 80.43\% (CS) and 79.91\%(CV) with 0.89 s for the stateful version.

arXiv:2006.07034공식 코드 · Papers with Code

Benchmarking Unsupervised Object Representations for Video Sequences

Perceiving the world in terms of objects and tracking them through time is a crucial prerequisite for reasoning and scene understanding. Recently, several methods have been proposed for unsupervised learning of object-centric representations. However, since these models were evaluated on different downstream tasks, it remains unclear how they compare in terms of basic perceptual abilities such as detection, figure-ground segmentation and tracking of objects. To close this gap, we design a benchmark with four data sets of varying complexity and seven additional test sets featuring challenging tracking scenarios relevant for natural videos. Using this benchmark, we compare the perceptual abilities of four object-centric approaches: ViMON, a video-extension of MONet, based on recurrent spatial attention, OP3, which exploits clustering via spatial mixture models, as well as TBA and SCALOR, which use explicit factorization via spatial transformers. Our results suggest that the architectures with unconstrained latent representations learn more powerful representations in terms of object detection, segmentation and tracking than the spatial transformer based architectures. We also observe that none of the methods are able to gracefully handle the most challenging tracking scenarios despite their synthetic nature, suggesting that our benchmark may provide fruitful guidance towards learning more robust object-centric video representations.

arXiv:2006.06158공식 코드 · Papers with Code

0-MMS: Zero-Shot Multi-Motion Segmentation With A Monocular Event Camera

Segmentation of moving objects in dynamic scenes is a key process in scene understanding for navigation tasks. Classical cameras suffer from motion blur in such scenarios rendering them effete. On the contrary, event cameras, because of their high temporal resolution and lack of motion blur, are tailor-made for this problem. We present an approach for monocular multi-motion segmentation, which combines bottom-up feature tracking and top-down motion compensation into a unified pipeline, which is the first of its kind to our knowledge. Using the events within a time-interval, our method segments the scene into multiple motions by splitting and merging. We further speed up our method by using the concept of motion propagation and cluster keyslices. The approach was successfully evaluated on both challenging real-world and synthetic scenarios from the EV-IMO, EED, and MOD datasets and outperformed the state-of-the-art detection rate by 12\%, achieving a new state-of-the-art average detection rate of 81.06%, 94.2% and 82.35% on the aforementioned datasets. To enable further research and systematic evaluation of multi-motion segmentation, we present and open-source a new dataset/benchmark called MOD++, which includes challenging sequences and extensive data stratification in-terms of camera and object motion, velocity magnitudes, direction, and rotational speeds.