Peng Liu

Image segmentation under low-light conditions is essential in real-world applications, such as autonomous driving and video surveillance systems. The recent Segment Anything Model (SAM) exhibits strong segmentation capability in various vision applications. However, its performance could be severely degraded under low-light conditions. On the other hand, multimodal information has been exploited to help models construct more comprehensive understanding of scenes under low-light conditions by providing complementary information (e.g., depth). Therefore, in this work, we present a pioneer attempt that elevates a unimodal vision foundation model (e.g., SAM) to a multimodal one, by efficiently integrating additional depth information under low-light conditions. To achieve that, we propose a novel method called Depth Perception SAM (DPSAM) based on the SAM framework. Specifically, we design a modality encoder to extract the depth information and the Depth Perception Layers (DPLs) for mutual feature refinement between RGB and depth features. The DPLs employ the cross-modal attention mechanism to mutually query effective information from both RGB and depth for the subsequent feature refinement. Thus, DPLs can effectively leverage the complementary information from depth to enrich the RGB representations and obtain comprehensive multimodal visual representations for segmenting anything in the dark. To this end, our DPSAM maximally maintains the instinct expertise of SAM for RGB image segmentation and further leverages on the strength of depth for enhanced segmenting anything capability, especially for cases that are likely to fail with RGB only (e.g., low-light or complex textures). As demonstrated by extensive experiments on four RGBD benchmark datasets, DPSAM clearly improves the performance for the segmenting anything performance in the dark, e.g., +12.90% mIoU and +16.23% mIoU on LLRGBD and DeLiVER, respectively.

Autonomous driving systems rely heavily on semantic segmentation models for accurate and safe decision-making. High segmentation performance in real-world urban scenes is crucial for autonomous vehicles, while substantial pixel-level labels are required during model training. Unsupervised domain adaptation (UDA) techniques are widely used to adapt the segmentation model trained on the synthetic data (i.e., source domain) to the real-world data (i.e., target domain) since obtaining pixel-level annotations is fairly easy in the synthetic environment. Recently, increasing UDA approaches promote cross-domain semantic segmentation (CDSS) by fusing the depth information into the RGB features. However, feature fusion does not necessarily eliminate the domain-specific components in the RGB features, which can result in the features still being influenced by domain-specific information. To address this, we propose a novel cross-modal attentive feature alignment (CAFA) framework for CDSS, which provides an explicit perspective of using depth information to align the main backbone RGB features of both domains in a nonadversarial manner. In particular, considering that the depth modality is less affected by the domain gap, we employ depth as an intermediate modality and align the RGB features by attending RGB features to the depth modality through constructing an auxiliary multimodal segmentation task. The state-of-the-art performance of our CAFA can be achieved on benchmark tasks, such as Synthia $\to$ Cityscapes and grand theft auto (GTA) $\to$ Cityscapes.

Synthetic data (i.e., source domain) have been widely adopted to improve the semantic segmentation performance for real-world images (i.e., target domain), since obtaining pixel-level annotations is fairly easy in the synthetic environment. Traditional domain adaptation methods normally focus on learning in the RGB modality only. We notice that the synthetic environment can generate depth information of semantic objects at almost no cost, while it is nontrivial to collect such information in the real-world scenario. In this case, we employ the depth information of synthetic data in this work to further boost the segmentation performance, and then transform the uni-modal problem into a multi-modal one. In this work, we focus on urban scene understanding and make a pioneer attempt on learning uni-modal feature representations for real-world images by mining from multi-modal knowledge of synthetic images with additional depth information. To this end, we propose a novel method called Multi-modal Domain Knowledge Transfer (MDKT), which transfers the multi-modal knowledge of the source domain to the uni-modal target domain through domain adaptation. In MDKT, we first employ the Cross-Modal Correlation (CMC) module to enhance the source features by fusing the RGB and depth information. Then, the uni-modal target domain feature and multi-modal source domain feature are aligned through the Modal-Imbalanced Adversarial Training (MIAT) strategy, which transfers the multi-modal knowledge to the uni-modal network in the target domain. We conduct extensive experiments on several benchmark settings for urban scene understanding. The promising results clearly show the effectiveness of our proposed MDKT approach.

A waveguide attenuator is developed for millimeter-wave system test applications. It is realized based on a standard WR6 rectangular waveguide; the substrate of the attenuator is quartz while the resistance element is the use of lossy material. A 15-dB attenuator is designed and validated from experiments.

In this paper, we study meandered topologies of a dual-mode loop resonator for microwave multiplexer applications. Compact channel filter structures that characterize controllable transmission zeros locating simultaneously at the lower stopband or upper stopband are, for the first time, investigated for such a kind of resonators and based on this, a microstrip triplexer is proposed. The studied triplexer uses one of the filter structures to realize the middle channel, and the rest structure to realize the upper and lower channels. The center frequencies of the three channels are 1.8 GHz, 2.0 GHz and 2.2 GHz, and the 3-dB fractional bandwidths are 3.6%, 3.6% and 4.1%, respectively. The developed triplexer features high channel-isolations of over 40 dB as a result of the controllable transmission zeros of the channel filters.

Aims: To apply a deep learning model for automatic localisation of the scleral spur (SS) in anterior segment optical coherence tomography (AS-OCT) images and compare the reproducibility of anterior chamber angle (ACA) width between deep learning located SS (DLLSS) and manually plotted SS (MPSS).Methods: In this multicentre, cross-sectional study, a test dataset comprising 5166 AS-OCT images from 287 eyes (116 healthy eyes with open angles and 171 eyes with primary angle-closure disease (PACD)) of 287 subjects were recruited from four ophthalmology clinics. Each eye was imaged twice by a swept-source AS-OCT (CASIA2, Tomey, Nagoya, Japan) in the same visit and one eye of each patient was randomly selected for measurements of ACA. The agreement between DLLSS and MPSS was assessed using the Euclidean distance (ED). The angle opening distance (AOD) of 750 µm (AOD750) and trabecular-iris space area (TISA) of 750 µm (TISA750) were calculated using the CASIA2 embedded software. The repeatability of ACA width was measured.Results: The mean age was 60.8±12.3 years (range: 30–85 years) for the normal group and 63.4±10.6 years (range: 40–91 years) for the PACD group. The mean difference in ED for SS localisation between DLLSS and MPSS was 66.50±20.54 µm and 84.78±28.33 µm for the normal group and the PACD group, respectively. The span of 95% limits of agreement between DLLSS and MPSS was 0.064 mm for AOD750 and 0.034 mm²for TISA750. The respective repeatability coefficients of AOD750 and TISA750 were 0.049 mm and 0.026 mm²for DLLSS, and 0.058 mm and 0.030 mm²for MPSS.Conclusion: DLLSS achieved comparable repeatability compared with MPSS for measurement of ACA.