We present ForceCtrl, a novel 3D hand raycasting technique that enhances pointing precision based on control-display (CD) gain controlled with user-defined pinch force. We introduce a target-agnostic approach for refining raycasting precision, overcoming limitations in human motor accuracy. User-defined pinch force, detected with surface electromyography (sEMG), enables users to easily activate or deactivate CD gain during interaction. We propose three CD gain strategies and compare them through target selection and placement tasks. Our system reduces selection errors, placement jitters, and user workload, especially for distant targets in high-difficulty tasks. These results highlight the effectiveness of applying CD gain to hand raycasting and demonstrate the potential of user-defined pinch force as a robust input modality for precise hand interaction in AR/VR.

Detection Transformers (DETR) have recently set new benchmarks in object detection. However, their performance in detecting rotated objects lags behind established oriented object detectors. Our analysis identifies a key observation: the boundary discontinuity and square-like problem in bipartite matching poses an issue with assigning appropriate ground truths to predictions, leading to duplicate low-confidence predictions. To address this, we introduce a Hausdorff distance-based cost for bipartite matching, which more accurately quantifies the discrepancy between predictions and ground truths. Additionally, we find that a static denoising approach impedes the training of rotated DETR, especially as the quality of the detector's predictions begins to exceed that of the noised ground truths. To overcome this, we propose an adaptive query denoising method that employs bipartite matching to selectively eliminate noised queries that detract from model improvement. When compared to models adopting a ResNet-50 backbone, our proposed model yields remarkable improvements, achieving$+4.18AP_{50}, +4.59AP_{50}, and+4.99AP_{50}$on DOTA-v2.0, DOTA-v1.5, and DIOR-R, respectively.

As understanding and monitoring global warming and heatwaves have become increasingly important, the demand for higher spatial and temporal resolution in satellite-observed Land Surface Temperature (LST) data has risen. LST derived from geostationary satellite observations plays a crucial role in temperature monitoring, offering high temporal resolution across wide areas. However, a primary limitation of geostationary satellite-derived LST products is their low spatial resolution. In this study, we aim to overcome this limitation by employing a deep learning-based super-resolution (SR) approach and propose a model calledReferences and Residual in Residual Blocks to Perform Super Resolution Network (3R-Net). This model incorporates terrain-guided reference images and residual blocks to enable more accurate super-resolution. By using LST products from GEO-KOMPSAT-2A (GK2A) as low-resolution input, visible channel imagery from GEO-KOMPSAT-2B (GK2B) as terrain-reflective high-resolution reference, and Landsat-8 LST products as high-resolution target images, our model effectively enhances the 2 km resolution of GK2A LST products to the 500 m resolution of Landsat-8 with improved spatial detail and accuracy. Unlike many previous studies relying on single-image super-resolution with synthetically downsampled inputs, our approach uses real-world LR inputs, HR targets, and reference images, making the learning process more realistic and practical. Experimental results show that, when clear guidance is provided through reference images, 3R-Net surpasses existing SR methods, achieving higher PSNR, SSIM, and lower RMSE while capturing critical spatial and temporal features, including surface characteristics and daily heating patterns. By integrating residual-in-residual blocks our model achieves a simple yet powerful enhancement in capturing fine-grained spatial and temporal patterns. These advancements suggest that 3R-Net can provide enhanced LST data crucial for climate research, environmental monitoring, and early warning systems.

We present PiMForce, a novel framework that enhances hand pressure estimation by leveraging 3D hand posture information to augment forearm surface electromyography (sEMG) signals. Our approach utilizes detailed spatial information from 3D hand poses in conjunction with dynamic muscle activity from sEMG to enable accurate and robust whole-hand pressure measurements under diverse hand-object interactions. We also developed a multimodal data collection system that combines a pressure glove, an sEMG armband, and a markerless finger-tracking module. We created a comprehensive dataset from 21 participants, capturing synchronized data of hand posture, sEMG signals, and exerted hand pressure across various hand postures and hand-object interaction scenarios using our collection system. Our framework enables precise hand pressure estimation in complex and natural interaction scenarios. Our approach substantially mitigates the limitations of traditional sEMG-based or vision-based methods by integrating 3D hand posture information with sEMG signals. Video demos, data, and code are available online.

For change detection in remote sensing, constructing a training dataset for deep learning models is difficult due to the requirements of bi-temporal supervision. To overcome this issue, single-temporal supervision which treats change labels as the difference of two semantic masks has been proposed. This novel method trains a change detector using two spatially unrelated images with corresponding semantic labels such as building. However, training on unpaired datasets could confuse the change detector in the case of pixels that are labeled unchanged but are visually significantly different. In order to maintain the visual similarity in unchanged area, in this paper, we emphasize that the change originates from the source image and show that manipulating the source image as an after-image is crucial to the performance of change detection. Extensive experiments demonstrate the importance of maintaining visual information between pre- and post-event images, and our method outperforms existing methods based on single-temporal supervision.

Adversarial examples, crafted by adding imperceptible perturbations to natural inputs, can easily fool deep neural networks (DNNs). One of the most successful methods for training adversarially robust DNNs is solving a nonconvex-nonconcave minimax problem with an adversarial training (AT) algorithm. However, among the many AT algorithms, only Dynamic AT (DAT) and You Only Propagate Once (YOPO) is guaranteed to converge to a stationary point with rate O(1/K^{1/2}). In this work, we generalize the stochastic primal-dual hybrid gradient algorithm to develop semi-implicit hybrid gradient methods (SI-HGs) for finding stationary points of nonconvex-nonconcave minimax problems. SI-HGs have the convergence rate O(1/K), which improves upon the rate O(1/K^{1/2}) of DAT and YOPO. We devise a practical variant of SI-HGs, and show that it outperforms other AT algorithms in terms of convergence speed and robustness.

In the training of deep learning models, how the model parameters are initialized greatly affects the model performance, sample efficiency, and convergence speed. Recently, representation learning for model initialization has been actively studied in the remote sensing field. In particular, the appearance characteristics of the imagery obtained using the synthetic aperture radar (SAR) sensor are quite different from those of general electro-optical (EO) images, and thus, representation learning is even more important in remote sensing domain. Motivated from contrastive multiview coding, we propose multimodal representation learning for SAR semantic segmentation. Unlike previous studies, our method jointly uses EO imagery, SAR imagery, and a label mask. Several experiments show that our approach is superior to the existing methods in model performance, sample efficiency, and convergence speed.

Road extraction from very high resolution (VHR) satellite images is one of the most important topics in the field of remote sensing. In this letter, we propose an efficient nonlocal LinkNet with nonlocal blocks (NLBs) that can grasp relations between global features. This enables each spatial feature point to refer to all other contextual information and results in more accurate road segmentation. In detail, our single model without any postprocessing like conditional random field (CRF) refinement performed better than any other published state-of-the-art ensemble model in the official DeepGlobe Challenge. Moreover, our nonlocal LinkNet (NL-LinkNet) beat the D-LinkNet, the winner of the DeepGlobe challenge (Demiret al., 2018), with 43% less parameters, less giga floating-point operations per seconds (GFLOPs), and shorter training convergence time. We also present empirical analyses on the proper usages of NLBs for the baseline model.

Partial label learning (PLL) is a class of weakly supervised learning where each training instance consists of a data and a set of candidate labels containing a unique ground truth label. To tackle this problem, a majority of current state-of-the-art methods employs either label disambiguation or averaging strategies. So far, PLL methods without such techniques have been considered impractical. In this paper, we challenge this view by revealing the hidden power of the oldest and naivest PLL method when it is instantiated with deep neural networks. Specifically, we show that, with deep neural networks, the naive model can achieve competitive performances against the other state-of-the-art methods, suggesting it as a strong baseline for PLL. We also address the question of how and why such a naive model works well with deep neural networks. Our empirical results indicate that deep neural networks trained on partially labeled examples generalize very well even in the over-parametrized regime and without label disambiguations or regularizations. We point out that existing learning theories on PLL are vacuous in the over-parametrized regime. Hence they cannot explain why the deep naive method works. We propose an alternative theory on how deep learning generalize in PLL problems.

In this paper, we propose a rotated bounding box based convolutional neural network (RBox-CNN) for arbitrary-oriented ship detection. RBox-CNN is an end-to-end model based on Faster R-CNN. The region proposal network generates proposals as the rotated bounding box, and then the rotation region-of-interest (RRoI) pooling layer is applied to extract region features corresponding the proposals. In addition, the diagonal region-of-interest (DRoI) pooling layer is applied simultaneously to extract context features and alleviate the problem of misalignment in RRoI pooling layer. To stably predict locations with the angle, we apply the regression of distance's projection in width/height. Experiments on HRSC2016 show that our model achieves state-of-the-art detection accuracy on ship detection. Furthermore, RBox-CNN achieves a significant improvement on DOTA for oriented general object detection in remote sensing images.