Delving Deep Into Label Smoothing Zhang, Chang-Bin; Jiang, Peng-Tao; Hou, Qibin ...
IEEE transactions on image processing,
2021, Volume:
30
Journal Article
Peer reviewed
Open access
Label smoothing is an effective regularization tool for deep neural networks (DNNs), which generates soft labels by applying a weighted average between the uniform distribution and the hard label. It ...is often used to reduce the overfitting problem of training DNNs and further improve classification performance. In this paper, we aim to investigate how to generate more reliable soft labels. We present an Online Label Smoothing (OLS) strategy, which generates soft labels based on the statistics of the model prediction for the target category. The proposed OLS constructs a more reasonable probability distribution between the target categories and non-target categories to supervise DNNs. Experiments demonstrate that based on the same classification models, the proposed approach can effectively improve the classification performance on CIFAR-100, ImageNet, and fine-grained datasets. Additionally, the proposed method can significantly improve the robustness of DNN models to noisy labels compared to current label smoothing approaches. The source code is available at our project page: https://mmcheng.net/ols/
Objective
To investigate the association between rheumatoid arthritis (RA) and deep neck infection (DNI).
Study Design
Retrospective cohort study.
Methods
Patients newly diagnosed with RA between ...2000 and 2011 were identified from the National Health Insurance Research Database in Taiwan. Moreover, patients without RA were randomly selected and matched at a 1:4 ratio by age, sex, urbanization level, income, and diabetes mellitus. The patients were followed up until death or the end of the study period (December 31, 2013). The primary outcome was the occurrence of DNI.
Results
In total, 30,207 patients with RA and 120,828 matched patients without RA were enrolled. Patients with RA had a significantly higher cumulative incidence of DNI than those without RA (P < 0.001). The adjusted Cox proportional hazard model demonstrated that RA was significantly associated with a higher incidence of DNI (hazard ratio: 2.80, 95% confidence interval: 2.26–3.46, P < 0.001). Therapeutic methods (surgical or nonsurgical) did not differ significantly between the patients with RA‐DNI and with non–RA‐DNI. Patients with RA‐DNI had higher rates of tracheostomy, mediastinitis, mediastinitis‐related mortality, and mortality than patients with non–RA‐DNI, although these differences were without statistical significance. RA patients receiving no therapy experienced higher rates of DNI compared with those receiving methotrexate alone, disease‐modifying antirheumatic drugs, or biologic therapies.
Conclusion
This study is the first to investigate the association between RA and DNI. We conclude RA is an independent predisposing factor for DNI.
Level of Evidence
4 Laryngoscope, 130:1402–1407, 2020
Continuous‐wave (CW) room‐temperature (RT) laser operation with low energy consumption is an ultimate goal for electrically driven lasers. A monolithically integrated perovskite laser in a chip‐level ...fiber scheme is ideal. However, because of the well‐recognized air and thermal instabilities of perovskites, laser action in a perovskite has mostly been limited to either pulsed or cryogenic‐temperature operations. Most CW laser operations at RT have had poor durability. Here, crystal fibers that have robust and high‐heat‐load nature are shown to be the key to enabling the first demonstration of ultralow‐threshold CW RT laser action in a compact, monolithic, and inexpensive crystal fiber/nanoperovskite hybrid architecture that is directly pumped with a 405 nm diode laser. Purcell‐enhanced light–matter coupling between the atomically smooth fiber microcavity and the perovskite nanocrystallites gain medium enables a high Q (≈1500) and a high β (0.31). This 762 nm laser outperforms previously reported structures with a record‐low threshold of 132 nW and an optical‐to‐optical slope conversion efficiency of 2.93%, and it delivers a stable output for CW and RT operation. These results represent a significant advancement toward monolithic all‐optical integration.
Purcell‐enhanced light–matter coupling between an atomically smooth fiber microcavity and perovskite gain nanocrystallites enables a high Q (≈1500) and a high β (0.31), outperforming previously reported structures with a 132 nW record‐low threshold. This is thought to be the first demonstration of monolithic integration of a perovskite laser into a compact and inexpensive fiber platform.
Global Contrast Based Salient Region Detection Cheng, Ming-Ming; Mitra, Niloy J.; Huang, Xiaolei ...
IEEE transactions on pattern analysis and machine intelligence,
2015-March-1, 2015-Mar, 2015-3-1, 20150301, Volume:
37, Issue:
3
Journal Article
Peer reviewed
Open access
Automatic estimation of salient object regions across images, without any prior assumption or knowledge of the contents of the corresponding scenes, enhances many computer vision and computer ...graphics applications. We introduce a regional contrast based salient object detection algorithm, which simultaneously evaluates global contrast differences and spatial weighted coherence scores. The proposed algorithm is simple, efficient, naturally multi-scale, and produces full-resolution, high-quality saliency maps. These saliency maps are further used to initialize a novel iterative version of GrabCut, namely SaliencyCut, for high quality unsupervised salient object segmentation. We extensively evaluated our algorithm using traditional salient object detection datasets, as well as a more challenging Internet image dataset. Our experimental results demonstrate that our algorithm consistently outperforms 15 existing salient object detection and segmentation methods, yielding higher precision and better recall rates. We also show that our algorithm can be used to efficiently extract salient object masks from Internet images, enabling effective sketch-based image retrieval (SBIR) via simple shape comparisons. Despite such noisy internet images, where the saliency regions are ambiguous, our saliency guided image retrieval achieves a superior retrieval rate compared with state-of-the-art SBIR methods, and additionally provides important target object region information.
Extracting roads from satellite imagery is a promising approach to update the dynamic changes of road networks efficiently and timely. However, it is challenging due to the occlusions caused by other ...objects and the complex traffic environment, the pixel-based methods often generate fragmented roads and fail to predict topological correctness. In this paper, motivated by the road shapes and connections in the graph network, we propose a connectivity attention network (CoANet) to jointly learn the segmentation and pair-wise dependencies. Since the strip convolution is more aligned with the shape of roads, which are long-span, narrow, and distributed continuously. We develop a strip convolution module (SCM) that leverages four strip convolutions to capture long-range context information from different directions and avoid interference from irrelevant regions. Besides, considering the occlusions in road regions caused by buildings and trees, a connectivity attention module (CoA) is proposed to explore the relationship between neighboring pixels. The CoA module incorporates the graphical information and enables the connectivity of roads are better preserved. Extensive experiments on the popular benchmarks (SpaceNet and DeepGlobe datasets) demonstrate that our proposed CoANet establishes new state-of-the-art results. The source code will be made publicly available at: https://mmcheng.net/coanet/ .
Predicting where people look in static scenes, a.k.a visual saliency, has received significant research interest recently. However, relatively less effort has been spent in understanding and modeling ...visual attention over dynamic scenes. This work makes three contributions to video saliency research. First, we introduce a new benchmark, called DHF1K (Dynamic Human Fixation 1K), for predicting fixations during dynamic scene free-viewing, which is a long-time need in this field. DHF1K consists of 1K high-quality elaborately-selected video sequences annotated by 17 observers using an eye tracker device. The videos span a wide range of scenes, motions, object types and backgrounds. Second, we propose a novel video saliency model, called ACLNet (Attentive CNN-LSTM Network), that augments the CNN-LSTM architecture with a supervised attention mechanism to enable fast end-to-end saliency learning. The attention mechanism explicitly encodes static saliency information, thus allowing LSTM to focus on learning a more flexible temporal saliency representation across successive frames. Such a design fully leverages existing large-scale static fixation datasets, avoids overfitting, and significantly improves training efficiency and testing performance. Third, we perform an extensive evaluation of the state-of-the-art saliency models on three datasets : DHF1K, Hollywood-2, and UCF sports. An attribute-based analysis of previous saliency models and cross-dataset generalization are also presented. Experimental results over more than 1.2K testing videos containing 400K frames demonstrate that ACLNet outperforms other contenders and has a fast processing speed (40 fps using a single GPU). Our code and all the results are available at https://github.com/wenguanwang/DHF1K.
The use of RGB-D information for salient object detection (SOD) has been extensively explored in recent years. However, relatively few efforts have been put toward modeling SOD in real-world human ...activity scenes with RGB-D. In this article, we fill the gap by making the following contributions to RGB-D SOD: 1) we carefully collect a new S al i ent P erson (SIP) data set that consists of ~1 K high-resolution images that cover diverse real-world scenes from various viewpoints, poses, occlusions, illuminations, and background s; 2) we conduct a large-scale (and, so far, the most comprehensive) benchmark comparing contemporary methods, which has long been missing in the field and can serve as a baseline for future research, and we systematically summarize 32 popular models and evaluate 18 parts of 32 models on seven data sets containing a total of about 97k images; and 3) we propose a simple general architecture, called deep depth-depurator network (D 3 Net). It consists of a depth depurator unit (DDU) and a three-stream feature learning module (FLM), which performs low-quality depth map filtering and cross-modal feature learning, respectively. These components form a nested structure and are elaborately designed to be learned jointly. D 3 Net exceeds the performance of any prior contenders across all five metrics under consideration, thus serving as a strong model to advance research in this field. We also demonstrate that D 3 Net can be used to efficiently extract salient object masks from real scenes, enabling effective background-changing application with a speed of 65 frames/s on a single GPU. All the saliency maps, our new SIP data set, the D 3 Net model, and the evaluation tools are publicly available at https://github.com/DengPingFan/D3NetBenchmark .
The tension biology of wound healing Harn, Hans I‐Chen; Ogawa, Rei; Hsu, Chao‐Kai ...
Experimental dermatology,
April 2019, Volume:
28, Issue:
4
Journal Article
Peer reviewed
Open access
Following skin wounding, the healing outcome can be: regeneration, repair with normal scar tissue, repair with hypertrophic scar tissue or the formation of keloids. The role of chemical factors in ...wound healing has been extensively explored, and while there is evidence suggesting the role of mechanical forces, its influence is much less well defined. Here, we provide a brief review on the recent progress of the role of mechanical force in skin wound healing by comparing laboratory mice, African spiny mice, fetal wound healing and adult scar keloid formation. A comparison across different species may provide insight into key regulators. Interestingly, some findings suggest tension can induce an immune response, and this provides a new link between mechanical and chemical forces. Clinically, manipulating skin tension has been demonstrated to be effective for scar prevention and treatment, but not for tissue regeneration. Utilising this knowledge, specialists may modulate regulatory factors and develop therapeutic strategies to reduce scar formation and promote regeneration.
In this paper, we present Vision Permutator, a conceptually simple and data efficient MLP-like architecture for visual recognition. By realizing the importance of the positional information carried ...by 2D feature representations, unlike recent MLP-like models that encode the spatial information along the flattened spatial dimensions, Vision Permutator separately encodes the feature representations along the height and width dimensions with linear projections. This allows Vision Permutator to capture long-range dependencies and meanwhile avoid the attention building process in transformers. The outputs are then aggregated in a mutually complementing manner to form expressive representations. We show that our Vision Permutators are formidable competitors to convolutional neural networks (CNNs) and vision transformers. Without the dependence on spatial convolutions or attention mechanisms, Vision Permutator achieves 81.5% top-1 accuracy on ImageNet without extra large-scale training data (e.g., ImageNet-22k) using only 25M learnable parameters, which is much better than most CNNs and vision transformers under the same model size constraint. When scaling up to 88M, it attains 83.2% top-1 accuracy, greatly improving the performance of recent state-of-the-art MLP-like networks for visual recognition. We hope this work could encourage research on rethinking the way of encoding spatial information and facilitate the development of MLP-like models. PyTorch/MindSpore/Jittor code is available at https://github.com/Andrew-Qibin/VisionPermutator .
The accurate assessment of liver fibrosis among hemodialysis patients with chronic hepatitis C (CHC) is important for both treatment and for follow up strategies. Applying the non-invasive methods in ...general population with viral hepatitis have been successful but the applicability of the aminotransferase/platelet ratio index (APRI) or the fibrosis-4 index (FIB-4) in hemodialysis patients need further evaluation.
We conducted a prospective, multi-center, uremic cohort to verify the applicability of APRI and FIB-4 in identifying liver fibrosis by reference with the standard transient elastography (TE) measures.
There were 116 CHC cases with valid TE were enrolled in our analysis. 46 cases (39.6%) were classified as F1, 35 cases (30.2%) as F2, 11 cases (9.5%) as F3, and 24 cases (20.7%) as F4, respectively. The traditional APRI and FIB-4 criteria did not correctly identify liver fibrosis. The optimal cut-off value of APRI was 0.28 and of FIB-4 was 1.91 to best excluding liver cirrhosis with AUC of 76% and 77%, respectively. The subgroup analysis showed that female CHC hemodialysis patients had better diagnostic accuracy with 74.1% by APRI. And CHC hemodialysis patients without hypertension had better diagnostic accuracy with 78.6% by FIB-4.
This study confirmed the traditional category level of APRI and FIB-4 were unable to identify liver fibrosis of CHC hemodialysis patients. With the adjusted cut-off value, APRI and FIB-4 still showed suboptimal diagnostic accuracy. Our results suggest the necessary of TE measures for liver fibrosis in the CHC uremic population.