PyTorchKR Papers

최신 논문

총 3,337편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2103.09030공식 코드 · Papers with Code

A Large-Scale Dataset for Benchmarking Elevator Button Segmentation and Character Recognition

Human activities are hugely restricted by COVID-19, recently. Robots that can conduct inter-floor navigation attract much public attention, since they can substitute human workers to conduct the service work. However, current robots either depend on human assistance or elevator retrofitting, and fully autonomous inter-floor navigation is still not available. As the very first step of inter-floor navigation, elevator button segmentation and recognition hold an important position. Therefore, we release the first large-scale publicly available elevator panel dataset in this work, containing 3,718 panel images with 35,100 button labels, to facilitate more powerful algorithms on autonomous elevator operation. Together with the dataset, a number of deep learning based implementations for button segmentation and recognition are also released to benchmark future methods in the community. The dataset will be available at \url{https://github.com/zhudelong/elevator_button_recognition

arXiv:2103.06447

Self-Supervised Motion Retargeting with Safety Guarantee

In this paper, we present self-supervised shared latent embedding (S3LE), a data-driven motion retargeting method that enables the generation of natural motions in humanoid robots from motion capture data or RGB videos. While it requires paired data consisting of human poses and their corresponding robot configurations, it significantly alleviates the necessity of time-consuming data-collection via novel paired data generating processes. Our self-supervised learning procedure consists of two steps: automatically generating paired data to bootstrap the motion retargeting, and learning a projection-invariant mapping to handle the different expressivity of humans and humanoid robots. Furthermore, our method guarantees that the generated robot pose is collision-free and satisfies position limits by utilizing nonparametric regression in the shared latent space. We demonstrate that our method can generate expressive robotic motions from both the CMU motion capture database and YouTube videos.

arXiv:2103.06417

3D Head-Position Prediction in First-Person View by Considering Head Pose for Human-Robot Eye Contact

For a humanoid robot to make eye contact and initiate communication with a person, it is necessary to estimate the person's head position. However, eye contact becomes difficult due to the mechanical delay of the robot when the person is moving. Owing to these issues, it is important to conduct a head-position prediction to mitigate the effect of the delay in the robot motion. Based on the fact that humans turn their heads before changing direction while walking, we hypothesized that the accuracy of three-dimensional (3D) head-position prediction from a first-person view can be improved by considering the head pose. We compared our method with a conventional Kalman filter-based approach, and found our method to be more accurate. The experiment results show that considering the head pose helps improve the accuracy of 3D head-position prediction.

arXiv:2103.06304

What is Multimodality?

The last years have shown rapid developments in the field of multimodal machine learning, combining e.g., vision, text or speech. In this position paper we explain how the field uses outdated definitions of multimodality that prove unfit for the machine learning era. We propose a new task-relative definition of (multi)modality in the context of multimodal machine learning that focuses on representations and information that are relevant for a given machine learning task. With our new definition of multimodality we aim to provide a missing foundation for multimodal research, an important component of language grounding and a crucial milestone towards NLU.

arXiv:2103.06879공식 코드 · Papers with Code

CoMoGAN: continuous model-guided image-to-image translation

CoMoGAN is a continuous GAN relying on the unsupervised reorganization of the target data on a functional manifold. To that matter, we introduce a new Functional Instance Normalization layer and residual mechanism, which together disentangle image content from position on target manifold. We rely on naive physics-inspired models to guide the training while allowing private model/translations features. CoMoGAN can be used with any GAN backbone and allows new types of image translation, such as cyclic image translation like timelapse generation, or detached linear translation. On all datasets, it outperforms the literature. Our code is available at http://github.com/cv-rits/CoMoGAN .

arXiv:2103.06725공식 코드 · Papers with Code

Duplex Contextual Relation Network for Polyp Segmentation

Polyp segmentation is of great importance in the early diagnosis and treatment of colorectal cancer. Since polyps vary in their shape, size, color, and texture, accurate polyp segmentation is very challenging. One promising way to mitigate the diversity of polyps is to model the contextual relation for each pixel such as using attention mechanism. However, previous methods only focus on learning the dependencies between the position within an individual image and ignore the contextual relation across different images. In this paper, we propose Duplex Contextual Relation Network (DCRNet) to capture both within-image and cross-image contextual relations. Specifically, we first design Interior Contextual-Relation Module to estimate the similarity between each position and all the positions within the same image. Then Exterior Contextual-Relation Module is incorporated to estimate the similarity between each position and the positions across different images. Based on the above two types of similarity, the feature at one position can be further enhanced by the contextual region embedding within and across images. To store the characteristic region embedding from all the images, a memory bank is designed and operates as a queue. Therefore, the proposed method can relate similar features even though they come from different images. We evaluate the proposed method on the EndoScene, Kvasir-SEG and the recently released large-scale PICCOLO dataset. Experimental results show that the proposed DCRNet outperforms the state-of-the-art methods in terms of the widely-used evaluation metrics.

arXiv:2103.06160

Using Cognitive Models to Train Warm Start Reinforcement Learning Agents for Human-Computer Interactions

Reinforcement learning (RL) agents in human-computer interactions applications require repeated user interactions before they can perform well. To address this "cold start" problem, we propose a novel approach of using cognitive models to pre-train RL agents before they are applied to real users. After briefly reviewing relevant cognitive models, we present our general methodological approach, followed by two case studies from our previous and ongoing projects. We hope this position paper stimulates conversations between RL, HCI, and cognitive science researchers in order to explore the full potential of the approach.

arXiv:2103.05267

Memory-Efficient, Limb Position-Aware Hand Gesture Recognition using Hyperdimensional Computing

Electromyogram (EMG) pattern recognition can be used to classify hand gestures and movements for human-machine interface and prosthetics applications, but it often faces reliability issues resulting from limb position change. One method to address this is dual-stage classification, in which the limb position is first determined using additional sensors to select between multiple position-specific gesture classifiers. While improving performance, this also increases model complexity and memory footprint, making a dual-stage classifier difficult to implement in a wearable device with limited resources. In this paper, we present sensor fusion of accelerometer and EMG signals using a hyperdimensional computing model to emulate dual-stage classification in a memory-efficient way. We demonstrate two methods of encoding accelerometer features to act as keys for retrieval of position-specific parameters from multiple models stored in superposition. Through validation on a dataset of 13 gestures in 8 limb positions, we obtain a classification accuracy of up to 93.34%, an improvement of 17.79% over using a model trained solely on EMG. We achieve this while only marginally increasing memory footprint over a single limb position model, requiring $8\times$ less memory than a traditional dual-stage classification architecture.

arXiv:2103.05270

PcmNet: Position-Sensitive Context Modeling Network for Temporal Action Localization

Temporal action localization is an important and challenging task that aims to locate temporal regions in real-world untrimmed videos where actions occur and recognize their classes. It is widely acknowledged that video context is a critical cue for video understanding, and exploiting the context has become an important strategy to boost localization performance. However, previous state-of-the-art methods focus more on exploring semantic context which captures the feature similarity among frames or proposals, and neglect positional context which is vital for temporal localization. In this paper, we propose a temporal-position-sensitive context modeling approach to incorporate both positional and semantic information for more precise action localization. Specifically, we first augment feature representations with directed temporal positional encoding, and then conduct attention-based information propagation, in both frame-level and proposal-level. Consequently, the generated feature representations are significantly empowered with the discriminative capability of encoding the position-aware context information, and thus benefit boundary detection and proposal evaluation. We achieve state-of-the-art performance on both two challenging datasets, THUMOS-14 and ActivityNet-1.3, demonstrating the effectiveness and generalization ability of our method.

arXiv:2103.04180공식 코드 · Papers with Code

Neural networks can understand compositional functions that humans do not, in the context of emergent communication

We show that it is possible to craft transformations that, applied to compositional grammars, result in grammars that neural networks can learn easily, but humans do not. This could explain the disconnect between current metrics of compositionality, that are arguably human-centric, and the ability of neural networks to generalize to unseen examples. We propose to use the transformations as a benchmark, ICY, which could be used to measure aspects of the compositional inductive bias of networks, and to search for networks with similar compositional inductive biases to humans. As an example of this approach, we propose a hierarchical model, HU-RNN, which shows an inductive bias towards position-independent, word-like groups of tokens.

arXiv:2103.04692

Semiotically-grounded distant viewing of diagrams: insights from two multimodal corpora

In this article, we bring together theories of multimodal communication and computational methods to study how primary school science diagrams combine multiple expressive resources. We position our work within the field of digital humanities, and show how annotations informed by multimodality research, which target expressive resources and discourse structure, allow imposing structure on the output of computational methods. We illustrate our approach by analysing two multimodal diagram corpora: the first corpus is intended to support research on automatic diagram processing, whereas the second is oriented towards studying diagrams as a mode of communication. Our results show that multimodally-informed annotations can bring out structural patterns in the diagrams, which also extend across diagrams that deal with different topics.

arXiv:2103.02845공식 코드 · Papers with Code

Camera-Space Hand Mesh Recovery via Semantic Aggregation and Adaptive 2D-1D Registration

Recent years have witnessed significant progress in 3D hand mesh recovery. Nevertheless, because of the intrinsic 2D-to-3D ambiguity, recovering camera-space 3D information from a single RGB image remains challenging. To tackle this problem, we divide camera-space mesh recovery into two sub-tasks, i.e., root-relative mesh recovery and root recovery. First, joint landmarks and silhouette are extracted from a single input image to provide 2D cues for the 3D tasks. In the root-relative mesh recovery task, we exploit semantic relations among joints to generate a 3D mesh from the extracted 2D cues. Such generated 3D mesh coordinates are expressed relative to a root position, i.e., wrist of the hand. In the root recovery task, the root position is registered to the camera space by aligning the generated 3D mesh back to 2D cues, thereby completing cameraspace 3D mesh recovery. Our pipeline is novel in that (1) it explicitly makes use of known semantic relations among joints and (2) it exploits 1D projections of the silhouette and mesh to achieve robust registration. Extensive experiments on popular datasets such as FreiHAND, RHD, and Human3.6M demonstrate that our approach achieves stateof-the-art performance on both root-relative mesh recovery and root recovery. Our code is publicly available at https://github.com/SeanChenxy/HandMesh.

arXiv:2103.03059

Sub-pixel face landmarks using heatmaps and a bag of tricks

Accurate face landmark localization is an essential part of face recognition, reconstruction and morphing. To accurately localize face landmarks, we present our heatmap regression approach. Each model consists of a MobileNetV2 backbone followed by several upscaling layers, with different tricks to optimize both performance and inference cost. We use five na\"ive face landmarks from a publicly available face detector to position and align the face instead of using the bounding box like traditional methods. Moreover, we show by adding random rotation, displacement and scaling -- after alignment -- that the model is more sensitive to the face position than orientation. We also show that it is possible to reduce the upscaling complexity by using a mixture of deconvolution and pixel-shuffle layers without impeding localization performance. We present our state-of-the-art face landmark localization model (ranking second on The 2nd Grand Challenge of 106-Point Facial Landmark Localization validation set). Finally, we test the effect on face recognition using these landmarks, using a publicly available model and benchmarks.

arXiv:2103.02528

Constraint Control of a Boom Crane System

Boom cranes are among the most used cranes to lift heavy loads. Although fairly simple mechanically, from the control viewpoint this kind of crane is a nonlinear underactuated system which presents several challenges, especially when con-trolled in the presence of constraints. To solve this problem, we propose an approach based on the Explicit Reference Governor (ERG), which does not require any online optimization, thus making it computationally inexpensive. The proposed control scheme is able to steer the crane to a desired position ensuring the respect of limited joint ranges, maximum oscillation angle, and the avoidance of static obstacles.

arXiv:2102.12505

Kernel-based framework to estimate deformations of pneumothorax lung using relative position of anatomical landmarks

In video-assisted thoracoscopic surgeries, successful procedures of nodule resection are highly dependent on the precise estimation of lung deformation between the inflated lung in the computed tomography (CT) images during preoperative planning and the deflated lung in the treatment views during surgery. Lungs in the pneumothorax state during surgery have a large volume change from normal lungs, making it difficult to build a mechanical model. The purpose of this study is to develop a deformation estimation method of the 3D surface of a deflated lung from a few partial observations. To estimate deformations for a largely deformed lung, a kernel regression-based solution was introduced. The proposed method used a few landmarks to capture the partial deformation between the 3D surface mesh obtained from preoperative CT and the intraoperative anatomical positions. The deformation for each vertex of the entire mesh model was estimated per-vertex as a relative position from the landmarks. The landmarks were placed in the anatomical position of the lung's outer contour. The method was applied on nine datasets of the left lungs of live Beagle dogs. Contrast-enhanced CT images of the lungs were acquired. The proposed method achieved a local positional error of vertices of 2.74 mm, Hausdorff distance of 6.11 mm, and Dice similarity coefficient of 0.94. Moreover, the proposed method could estimate lung deformations from a small number of training cases and a small observation area. This study contributes to the data-driven modeling of pneumothorax deformation of the lung.

arXiv:2102.11588공식 코드 · Papers with Code

Data Fusion for Audiovisual Speaker Localization: Extending Dynamic Stream Weights to the Spatial Domain

Estimating the positions of multiple speakers can be helpful for tasks like automatic speech recognition or speaker diarization. Both applications benefit from a known speaker position when, for instance, applying beamforming or assigning unique speaker identities. Recently, several approaches utilizing acoustic signals augmented with visual data have been proposed for this task. However, both the acoustic and the visual modality may be corrupted in specific spatial regions, for instance due to poor lighting conditions or to the presence of background noise. This paper proposes a novel audiovisual data fusion framework for speaker localization by assigning individual dynamic stream weights to specific regions in the localization space. This fusion is achieved via a neural network, which combines the predictions of individual audio and video trackers based on their time- and location-dependent reliability. A performance evaluation using audiovisual recordings yields promising results, with the proposed fusion approach outperforming all baseline models.

arXiv:2102.10290

Contextual Argument Component Classification for Class Discussions

Argument mining systems often consider contextual information, i.e. information outside of an argumentative discourse unit, when trained to accomplish tasks such as argument component identification, classification, and relation extraction. However, prior work has not carefully analyzed the utility of different contextual properties in context-aware models. In this work, we show how two different types of contextual information, local discourse context and speaker context, can be incorporated into a computational model for classifying argument components in multi-party classroom discussions. We find that both context types can improve performance, although the improvements are dependent on context size and position.

arXiv:2102.11090

Position Information in Transformers: An Overview

Transformers are arguably the main workhorse in recent Natural Language Processing research. By definition a Transformer is invariant with respect to reordering of the input. However, language is inherently sequential and word order is essential to the semantics and syntax of an utterance. In this article, we provide an overview and theoretical comparison of existing methods to incorporate position information into Transformer models. The objectives of this survey are to (1) showcase that position information in Transformer is a vibrant and extensive research area; (2) enable the reader to compare existing methods by providing a unified notation and systematization of different approaches along important model dimensions; (3) indicate what characteristics of an application should be taken into account when selecting a position encoding; (4) provide stimuli for future research.

arXiv:2102.10882공식 코드 · Papers with Code

Conditional Positional Encodings for Vision Transformers

We propose a conditional positional encoding (CPE) scheme for vision Transformers. Unlike previous fixed or learnable positional encodings, which are pre-defined and independent of input tokens, CPE is dynamically generated and conditioned on the local neighborhood of the input tokens. As a result, CPE can easily generalize to the input sequences that are longer than what the model has ever seen during training. Besides, CPE can keep the desired translation-invariance in the image classification task, resulting in improved performance. We implement CPE with a simple Position Encoding Generator (PEG) to get seamlessly incorporated into the current Transformer framework. Built on PEG, we present Conditional Position encoding Vision Transformer (CPVT). We demonstrate that CPVT has visually similar attention maps compared to those with learned positional encodings and delivers outperforming results. Our code is available at https://github.com/Meituan-AutoML/CPVT .

arXiv:2102.10084공식 코드 · Papers with Code

Hate-Alert@DravidianLangTech-EACL2021: Ensembling strategies for Transformer-based Offensive language Detection

Social media often acts as breeding grounds for different forms of offensive content. For low resource languages like Tamil, the situation is more complex due to the poor performance of multilingual or language-specific models and lack of proper benchmark datasets. Based on this shared task, Offensive Language Identification in Dravidian Languages at EACL 2021, we present an exhaustive exploration of different transformer models, We also provide a genetic algorithm technique for ensembling different models. Our ensembled models trained separately for each language secured the first position in Tamil, the second position in Kannada, and the first position in Malayalam sub-tasks. The models and codes are provided.