PyTorchKR Papers

최신 논문

총 12,272편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2210.03686공식 코드 · Papers with Code

Humans need not label more humans: Occlusion Copy & Paste for Occluded Human Instance Segmentation

Modern object detection and instance segmentation networks stumble when picking out humans in crowded or highly occluded scenes. Yet, these are often scenarios where we require our detectors to work well. Many works have approached this problem with model-centric improvements. While they have been shown to work to some extent, these supervised methods still need sufficient relevant examples (i.e. occluded humans) during training for the improvements to be maximised. In our work, we propose a simple yet effective data-centric approach, Occlusion Copy & Paste, to introduce occluded examples to models during training - we tailor the general copy & paste augmentation approach to tackle the difficult problem of same-class occlusion. It improves instance segmentation performance on occluded scenarios for "free" just by leveraging on existing large-scale datasets, without additional data or manual labelling needed. In a principled study, we show whether various proposed add-ons to the copy & paste augmentation indeed contribute to better performance. Our Occlusion Copy & Paste augmentation is easily interoperable with any models: by simply applying it to a recent generic instance segmentation model without explicit model architectural design to tackle occlusion, we achieve state-of-the-art instance segmentation performance on the very challenging OCHuman dataset. Source code is available at https://github.com/levan92/occlusion-copy-paste.

arXiv:2210.03566공식 코드 · Papers with Code

Automated segmentation and morphological characterization of placental histology images based on a single labeled image

In this study, a novel method of data augmentation has been presented for the segmentation of placental histological images when the labeled data are scarce. This method generates new realizations of the placenta intervillous morphology while maintaining the general textures and orientations. As a result, a diversified artificial dataset of images is generated that can be used for training deep learning segmentation models. We have observed that on average the presented method of data augmentation led to a 42% decrease in the binary cross-entropy loss of the validation dataset compared to the common approach in the literature. Additionally, the morphology of the intervillous space is studied under the effect of the proposed image reconstruction technique, and the diversity of the artificially generated population is quantified. Due to the high resemblance of the generated images to the real ones, the applications of the proposed method may not be limited to placental histological images, and it is recommended that other types of tissues be investigated in future studies.

arXiv:2210.02509

Revisiting Syllables in Language Modelling and their Application on Low-Resource Machine Translation

Language modelling and machine translation tasks mostly use subword or character inputs, but syllables are seldom used. Syllables provide shorter sequences than characters, require less-specialised extracting rules than morphemes, and their segmentation is not impacted by the corpus size. In this study, we first explore the potential of syllables for open-vocabulary language modelling in 21 languages. We use rule-based syllabification methods for six languages and address the rest with hyphenation, which works as a syllabification proxy. With a comparable perplexity, we show that syllables outperform characters and other subwords. Moreover, we study the importance of syllables on neural machine translation for a non-related and low-resource language-pair (Spanish--Shipibo-Konibo). In pairwise and multilingual systems, syllables outperform unsupervised subwords, and further morphological segmentation methods, when translating into a highly synthetic language with a transparent orthography (Shipibo-Konibo). Finally, we perform some human evaluation, and discuss limitations and opportunities.

arXiv:2210.02834공식 코드 · Papers with Code

Robust Double-Encoder Network for RGB-D Panoptic Segmentation

Perception is crucial for robots that act in real-world environments, as autonomous systems need to see and understand the world around them to act properly. Panoptic segmentation provides an interpretation of the scene by computing a pixelwise semantic label together with instance IDs. In this paper, we address panoptic segmentation using RGB-D data of indoor scenes. We propose a novel encoder-decoder neural network that processes RGB and depth separately through two encoders. The features of the individual encoders are progressively merged at different resolutions, such that the RGB features are enhanced using complementary depth information. We propose a novel merging approach called ResidualExcite, which reweighs each entry of the feature map according to its importance. With our double-encoder architecture, we are robust to missing cues. In particular, the same model can train and infer on RGB-D, RGB-only, and depth-only input data, without the need to train specialized models. We evaluate our method on publicly available datasets and show that our approach achieves superior results compared to other common approaches for panoptic segmentation.

arXiv:2210.03105공식 코드 · Papers with Code

Mask3D: Mask Transformer for 3D Semantic Instance Segmentation

Modern 3D semantic instance segmentation approaches predominantly rely on specialized voting mechanisms followed by carefully designed geometric clustering techniques. Building on the successes of recent Transformer-based methods for object detection and image segmentation, we propose the first Transformer-based approach for 3D semantic instance segmentation. We show that we can leverage generic Transformer building blocks to directly predict instance masks from 3D point clouds. In our model called Mask3D each object instance is represented as an instance query. Using Transformer decoders, the instance queries are learned by iteratively attending to point cloud features at multiple scales. Combined with point features, the instance queries directly yield all instance masks in parallel. Mask3D has several advantages over current state-of-the-art approaches, since it neither relies on (1) voting schemes which require hand-selected geometric properties (such as centers) nor (2) geometric grouping mechanisms requiring manually-tuned hyper-parameters (e.g. radii) and (3) enables a loss that directly optimizes instance masks. Mask3D sets a new state-of-the-art on ScanNet test (+6.2 mAP), S3DIS 6-fold (+10.1 mAP), STPLS3D (+11.2 mAP) and ScanNet200 test (+12.4 mAP).

arXiv:2210.02992공식 코드 · Papers with Code

COVID-19 Detection Using Segmentation, Region Extraction and Classification Pipeline

The main purpose of this study is to develop a pipeline for COVID-19 detection from a big and challenging database of Computed Tomography (CT) images. The proposed pipeline includes a segmentation part, a lung extraction part, and a classifier part. Optional slice removal techniques after UNet-based segmentation of slices were also tried. The methodologies tried in the segmentation part are traditional segmentation methods as well as UNet-based methods. In the classification part, a Convolutional Neural Network (CNN) was used to take the final diagnosis decisions. In terms of the results: in the segmentation part, the proposed segmentation methods show high dice scores on a publicly available dataset. In the classification part, the results were compared at slice-level and at patient-level as well. At slice-level, methods were compared and showed high validation accuracy indicating efficiency in predicting 2D slices. At patient level, the proposed methods were also compared in terms of validation accuracy and macro F1 score on the validation set. The dataset used for classification is COV-19CT Database. The method proposed here showed improvement from our precious results on the same dataset. In Conclusion, the improved work in this paper has potential clinical usages for COVID-19 detection and diagnosis via CT images. The code is on github at https://github.com/IDU-CVLab/COV19D_3rd

arXiv:2210.02057공식 코드 · Papers with Code

Comparing Hysteresis Comparator and RMS Threshold Methods for Automatic Single Cough Segmentations

Research on diagnosing diseases based on voice signals currently are rapidly increasing, including cough-related diseases. When training the cough sound signals into deep learning models, it is necessary to have a standard input by segmenting several cough signals into individual cough signals. Previous research has been developed to segment cough signals from non-cough signals. This research evaluates the segmentation methods of several cough signals from a single audio file into several single-cough signals. We evaluate three different methods employing manual segmentation as a baseline and automatic segmentation. The results by two automatic segmentation methods obtained precisions of 73% and 70% compared to 49% by manual segmentation. The agreements of listening tests to count the number of correct single-cough segmentations show fair and moderate correlations for automatic segmentation methods and are comparable with manual segmentation.

arXiv:2210.02033공식 코드 · Papers with Code

Learning Video-independent Eye Contact Segmentation from In-the-Wild Videos

Human eye contact is a form of non-verbal communication and can have a great influence on social behavior. Since the location and size of the eye contact targets vary across different videos, learning a generic video-independent eye contact detector is still a challenging task. In this work, we address the task of one-way eye contact detection for videos in the wild. Our goal is to build a unified model that can identify when a person is looking at his gaze targets in an arbitrary input video. Considering that this requires time-series relative eye movement information, we propose to formulate the task as a temporal segmentation. Due to the scarcity of labeled training data, we further propose a gaze target discovery method to generate pseudo-labels for unlabeled videos, which allows us to train a generic eye contact segmentation model in an unsupervised way using in-the-wild videos. To evaluate our proposed approach, we manually annotated a test dataset consisting of 52 videos of human conversations. Experimental results show that our eye contact segmentation model outperforms the previous video-dependent eye contact detector and can achieve 71.88% framewise accuracy on our annotated test set. Our code and evaluation dataset are available at https://github.com/ut-vision/Video-Independent-ECS.

arXiv:2210.02030

Point Cloud Recognition with Position-to-Structure Attention Transformers

In this paper, we present Position-to-Structure Attention Transformers (PS-Former), a Transformer-based algorithm for 3D point cloud recognition. PS-Former deals with the challenge in 3D point cloud representation where points are not positioned in a fixed grid structure and have limited feature description (only 3D coordinates ($x, y, z$) for scattered points). Existing Transformer-based architectures in this domain often require a pre-specified feature engineering step to extract point features. Here, we introduce two new aspects in PS-Former: 1) a learnable condensation layer that performs point downsampling and feature extraction; and 2) a Position-to-Structure Attention mechanism that recursively enriches the structural information with the position attention branch. Compared with the competing methods, while being generic with less heuristics feature designs, PS-Former demonstrates competitive experimental results on three 3D point cloud tasks including classification, part segmentation, and scene segmentation.

arXiv:2210.02025공식 코드 · Papers with Code

GMMSeg: Gaussian Mixture based Generative Semantic Segmentation Models

Prevalent semantic segmentation solutions are, in essence, a dense discriminative classifier of p(class|pixel feature). Though straightforward, this de facto paradigm neglects the underlying data distribution p(pixel feature|class), and struggles to identify out-of-distribution data. Going beyond this, we propose GMMSeg, a new family of segmentation models that rely on a dense generative classifier for the joint distribution p(pixel feature,class). For each class, GMMSeg builds Gaussian Mixture Models (GMMs) via Expectation-Maximization (EM), so as to capture class-conditional densities. Meanwhile, the deep dense representation is end-to-end trained in a discriminative manner, i.e., maximizing p(class|pixel feature). This endows GMMSeg with the strengths of both generative and discriminative models. With a variety of segmentation architectures and backbones, GMMSeg outperforms the discriminative counterparts on three closed-set datasets. More impressively, without any modification, GMMSeg even performs well on open-world datasets. We believe this work brings fundamental insights into the related fields.

arXiv:2210.01891

Adaptively Weighted Data Augmentation Consistency Regularization for Robust Optimization under Concept Shift

Concept shift is a prevailing problem in natural tasks like medical image segmentation where samples usually come from different subpopulations with variant correlations between features and labels. One common type of concept shift in medical image segmentation is the "information imbalance" between label-sparse samples with few (if any) segmentation labels and label-dense samples with plentiful labeled pixels. Existing distributionally robust algorithms have focused on adaptively truncating/down-weighting the "less informative" (i.e., label-sparse in our context) samples. To exploit data features of label-sparse samples more efficiently, we propose an adaptively weighted online optimization algorithm -- AdaWAC -- to incorporate data augmentation consistency regularization in sample reweighting. Our method introduces a set of trainable weights to balance the supervised loss and unsupervised consistency regularization of each sample separately. At the saddle point of the underlying objective, the weights assign label-dense samples to the supervised loss and label-sparse samples to the unsupervised consistency regularization. We provide a convergence guarantee by recasting the optimization as online mirror descent on a saddle point problem. Our empirical results demonstrate that AdaWAC not only enhances the segmentation performance and sample efficiency but also improves the robustness to concept shift on various medical image segmentation tasks with different UNet-style backbones.

arXiv:2210.01884

Self-supervised Pre-training for Semantic Segmentation in an Indoor Scene

The ability to endow maps of indoor scenes with semantic information is an integral part of robotic agents which perform different tasks such as target driven navigation, object search or object rearrangement. The state-of-the-art methods use Deep Convolutional Neural Networks (DCNNs) for predicting semantic segmentation of an image as useful representation for these tasks. The accuracy of semantic segmentation depends on the availability and the amount of labeled data from the target environment or the ability to bridge the domain gap between test and training environment. We propose RegConsist, a method for self-supervised pre-training of a semantic segmentation model, exploiting the ability of the agent to move and register multiple views in the novel environment. Given the spatial and temporal consistency cues used for pixel level data association, we use a variant of contrastive learning to train a DCNN model for predicting semantic segmentation from RGB views in the target environment. The proposed method outperforms models pre-trained on ImageNet and achieves competitive performance when using models that are trained for exactly the same task but on a different dataset. We also perform various ablation studies to analyze and demonstrate the efficacy of our proposed method.

arXiv:2210.02416

A deep learning model for brain vessel segmentation in 3DRA with arteriovenous malformations

Segmentation of brain arterio-venous malformations (bAVMs) in 3D rotational angiographies (3DRA) is still an open problem in the literature, with high relevance for clinical practice. While deep learning models have been applied for segmenting the brain vasculature in these images, they have never been used in cases with bAVMs. This is likely caused by the difficulty to obtain sufficiently annotated data to train these approaches. In this paper we introduce a first deep learning model for blood vessel segmentation in 3DRA images of patients with bAVMs. To this end, we densely annotated 5 3DRA volumes of bAVM cases and used these to train two alternative 3DUNet-based architectures with different segmentation objectives. Our results show that the networks reach a comprehensive coverage of relevant structures for bAVM analysis, much better than what is obtained using standard methods. This is promising for achieving a better topological and morphological characterisation of the bAVM structures of interest. Furthermore, the models have the ability to segment venous structures even when missing in the ground truth labelling, which is relevant for planning interventional treatments. Ultimately, these results could be used as more reliable first initial guesses, alleviating the cumbersome task of creating manual labels.

arXiv:2210.02270공식 코드 · Papers with Code

Weak-shot Semantic Segmentation via Dual Similarity Transfer

Semantic segmentation is an important and prevalent task, but severely suffers from the high cost of pixel-level annotations when extending to more classes in wider applications. To this end, we focus on the problem named weak-shot semantic segmentation, where the novel classes are learnt from cheaper image-level labels with the support of base classes having off-the-shelf pixel-level labels. To tackle this problem, we propose SimFormer, which performs dual similarity transfer upon MaskFormer. Specifically, MaskFormer disentangles the semantic segmentation task into two sub-tasks: proposal classification and proposal segmentation for each proposal. Proposal segmentation allows proposal-pixel similarity transfer from base classes to novel classes, which enables the mask learning of novel classes. We also learn pixel-pixel similarity from base classes and distill such class-agnostic semantic similarity to the semantic masks of novel classes, which regularizes the segmentation model with pixel-level semantic relationship across images. In addition, we propose a complementary loss to facilitate the learning of novel classes. Comprehensive experiments on the challenging COCO-Stuff-10K and ADE20K datasets demonstrate the effectiveness of our method. Codes are available at https://github.com/bcmi/SimFormer-Weak-Shot-Semantic-Segmentation.

arXiv:2210.02203

PriorNet: lesion segmentation in PET-CT including prior tumor appearance information

Tumor segmentation in PET-CT images is challenging due to the dual nature of the acquired information: low metabolic information in CT and low spatial resolution in PET. U-Net architecture is the most common and widely recognized approach when developing a fully automatic image segmentation method in the medical field. We proposed a two-step approach, aiming to refine and improve the segmentation performances of tumoral lesions in PET-CT. The first step generates a prior tumor appearance map from the PET-CT volumes, regarded as prior tumor information. The second step, consisting of a standard U-Net, receives the prior tumor appearance map and PET-CT images to generate the lesion mask. We evaluated the method on the 1014 cases available for the AutoPET 2022 challenge, and the results showed an average Dice score of 0.701 on the positive cases.

arXiv:2210.02088공식 코드 · Papers with Code

WUDA: Unsupervised Domain Adaptation Based on Weak Source Domain Labels

Unsupervised domain adaptation (UDA) for semantic segmentation addresses the cross-domain problem with fine source domain labels. However, the acquisition of semantic labels has always been a difficult step, many scenarios only have weak labels (e.g. bounding boxes). For scenarios where weak supervision and cross-domain problems coexist, this paper defines a new task: unsupervised domain adaptation based on weak source domain labels (WUDA). To explore solutions for this task, this paper proposes two intuitive frameworks: 1) Perform weakly supervised semantic segmentation in the source domain, and then implement unsupervised domain adaptation; 2) Train an object detection model using source domain data, then detect objects in the target domain and implement weakly supervised semantic segmentation. We observe that the two frameworks behave differently when the datasets change. Therefore, we construct dataset pairs with a wide range of domain shifts and conduct extended experiments to analyze the impact of different domain shifts on the two frameworks. In addition, to measure domain shift, we apply the metric representation shift to urban landscape image segmentation for the first time. The source code and constructed datasets are available at \url{https://github.com/bupt-ai-cz/WUDA}.

논문 ID: document/9888149

A Novel Unsupervised Domain Adaption Method for Depth-Guided Semantic Segmentation Using Coarse-to-Fine Alignment

Domain adaptation methods in machine learning deal with the domain shift issue by aligning source and target data representation. This paper proposes a novel domain adaptation method for semantic segmentation that exploits the Fourier transform on chromatic space to improve the quality of style transfer, and generates pseudo-labels for self-training by combining the results from different teachers obtained at different rounds of self-training. Our method also applies class-level adversarial learning to achieve a more fine-grained alignment between the two domains, and a late fusion with a depth-estimation model to improve its segmentation outputs. Experiments show that our method yields superior performance in terms of accuracy compared to other existing state-of-the-art methods.

arXiv:2210.01467

An Anatomy-aware Framework for Automatic Segmentation of Parotid Tumor from Multimodal MRI

Magnetic Resonance Imaging (MRI) plays an important role in diagnosing the parotid tumor, where accurate segmentation of tumors is highly desired for determining appropriate treatment plans and avoiding unnecessary surgery. However, the task remains nontrivial and challenging due to ambiguous boundaries and various sizes of the tumor, as well as the presence of a large number of anatomical structures around the parotid gland that are similar to the tumor. To overcome these problems, we propose a novel anatomy-aware framework for automatic segmentation of parotid tumors from multimodal MRI. First, a Transformer-based multimodal fusion network PT-Net is proposed in this paper. The encoder of PT-Net extracts and fuses contextual information from three modalities of MRI from coarse to fine, to obtain cross-modality and multi-scale tumor information. The decoder stacks the feature maps of different modalities and calibrates the multimodal information using the channel attention mechanism. Second, considering that the segmentation model is prone to be disturbed by similar anatomical structures and make wrong predictions, we design anatomy-aware loss. By calculating the distance between the activation regions of the prediction segmentation and the ground truth, our loss function forces the model to distinguish similar anatomical structures with the tumor and make correct predictions. Extensive experiments with MRI scans of the parotid tumor showed that our PT-Net achieved higher segmentation accuracy than existing networks. The anatomy-aware loss outperformed state-of-the-art loss functions for parotid tumor segmentation. Our framework can potentially improve the quality of preoperative diagnosis and surgery planning of parotid tumors.

arXiv:2210.01315

Concurrent build direction, part segmentation, and topology optimization for additive manufacturing using neural networks

We propose a neural network-based approach to topology optimization that aims to reduce the use of support structures in additive manufacturing. Our approach uses a network architecture that allows the simultaneous determination of an optimized: (1) part segmentation, (2) the topology of each part, and (3) the build direction of each part that collectively minimize the amount of support structure. Through training, the network learns a material density and segment classification in the continuous 3D space. Given a problem domain with prescribed load and displacement boundary conditions, the neural network takes as input 3D coordinates of the voxelized domain as training samples and outputs a continuous density field. Since the neural network for topology optimization learns the density distribution field, analytical solutions to the density gradient can be obtained from the input-output relationship of the neural network. We demonstrate our approach on several compliance minimization problems with volume fraction constraints, where support volume minimization is added as an additional criterion to the objective function. We show that simultaneous optimization of part segmentation along with the topology and print angle optimization further reduces the support structure, compared to a combined print angle and topology optimization without segmentation.

arXiv:2210.01595공식 코드 · Papers with Code

FreDSNet: Joint Monocular Depth and Semantic Segmentation with Fast Fourier Convolutions

In this work we present FreDSNet, a deep learning solution which obtains semantic 3D understanding of indoor environments from single panoramas. Omnidirectional images reveal task-specific advantages when addressing scene understanding problems due to the 360-degree contextual information about the entire environment they provide. However, the inherent characteristics of the omnidirectional images add additional problems to obtain an accurate detection and segmentation of objects or a good depth estimation. To overcome these problems, we exploit convolutions in the frequential domain obtaining a wider receptive field in each convolutional layer. These convolutions allow to leverage the whole context information from omnidirectional images. FreDSNet is the first network that jointly provides monocular depth estimation and semantic segmentation from a single panoramic image exploiting fast Fourier convolutions. Our experiments show that FreDSNet has similar performance as specific state of the art methods for semantic segmentation and depth estimation. FreDSNet code is publicly available in https://github.com/Sbrunoberenguel/FreDSNet