Semantic segmentation of remote sensing (RS) images is vital in various practical applications, including urban construction planning, natural disaster monitoring, and land resources investigation. However, RS images are captured by airplanes or satellites at high altitudes and long distances, resulting in ground objects of the same category being scattered in various corners of the image. Moreover, objects of different sizes appear simultaneously in RS images. For example, some objects occupy a large area in urban scenes, while others only have small regions. Technically, the above two universal situations pose significant challenges to the segmentation with a high quality for RS images. Based on these observations, this paper proposes a Mask2Former with an improved query (IQ2Former) for this task. The fundamental motivation behind the IQ2Former is to enhance the capability of the query of Mask2Former by exploiting the characteristics of RS images well. First, we propose the Query Scenario Module (QSM), which aims to learn and group the queries from feature maps, allowing the selection of distinct scenarios such as the urban and rural areas, building clusters, and parking lots. Second, we design the query position module (QPM), which is developed to assign the image position information to each query without increasing the number of parameters, thereby enhancing the model’s sensitivity to small targets in complex scenarios. Finally, we propose the query attention module (QAM), which is constructed to leverage the characteristics of query attention to extract valuable features from the preceding queries. Being positioned between the duplicated transformer decoder layers, QAM ensures the comprehensive utilization of the supervisory information and the exploitation of those fine-grained details. Architecturally, the QSM, QPM, and QAM as well as an end-to-end model are assembled to achieve high-quality semantic segmentation. In comparison to the classical or state-of-the-art models (FCN, PSPNet, DeepLabV3+, OCRNet, UPerNet, MaskFormer, Mask2Former), IQ2Former has demonstrated exceptional performance across three publicly challenging remote-sensing image datasets, 83.59 mIoU on the Vaihingen dataset, 87.89 mIoU on Potsdam dataset, and 56.31 mIoU on LoveDA dataset. Additionally, overall accuracy, ablation experiment, and visualization segmentation results all indicate IQ2Former validity.

Semantic segmentation of remote-sensing (RS) images is one of the most fundamental tasks in the understanding of a remote-sensing scene. However, high-resolution RS images contain plentiful detailed information about ground objects, which scatter everywhere spatially and have variable sizes, styles, and visual appearances. Due to the high similarity between classes and diversity within classes, it is challenging to obtain satisfactory and accurate semantic segmentation results. This paper proposes a Dynamic High-Resolution Network (DyHRNet) to solve this problem. Our proposed network takes HRNet as a super-architecture, aiming to leverage the important connections and channels by further investigating the parallel streams at different resolution representations of the original HRNet. The learning task is conducted under the framework of a neural architecture search (NAS) and channel-wise attention module. Specifically, the Accelerated Proximal Gradient (APG) algorithm is introduced to iteratively solve the sparse regularization subproblem from the perspective of neural architecture search. In this way, valuable connections are selected for cross-resolution feature fusion. In addition, a channel-wise attention module is designed to weight the channel contributions for feature aggregation. Finally, DyHRNet fully realizes the dynamic advantages of data adaptability by combining the APG algorithm and channel-wise attention module simultaneously. Compared with nine classical or state-of-the-art models (FCN, UNet, PSPNet, DeepLabV3+, OCRNet, SETR, SegFormer, HRNet+FCN, and HRNet+OCR), DyHRNet has shown high performance on three public challenging RS image datasets (Vaihingen, Potsdam, and LoveDA). Furthermore, the visual segmentation results, the learned structures, the iteration process analysis, and the ablation study all demonstrate the effectiveness of our proposed model.

Open-vocabulary semantic segmentation (OVSS) aims to segment and recognize objects universally. Trained on extensive high-quality segmentation data, the segment anything model (SAM) has demonstrated remarkable universal segmentation capabilities, offering valuable support for OVSS. Although previous methods have made progress in leveraging SAM for OVSS, there are still some challenges: (1) SAM's tendency to over-segment and (2) hard combinations between fixed masks and labels. This paper introduces a novel mask-injected framework, SAM-MI, which effectively integrates SAM with OVSS models to address these challenges. Initially, SAM-MI employs a Text-guided Sparse Point Prompter to sample sparse prompts for SAM instead of previous dense grid-like prompts, thus significantly accelerating the mask generation process. The framework then introduces Shallow Mask Aggregation (SMAgg) to merge partial masks to mitigate the SAM's over-segmentation issue. Finally, Decoupled Mask Injection (DMI) incorporates SAM-generated masks for guidance at low-frequency and high-frequency separately, rather than directly combining them with labels. Extensive experiments on multiple benchmarks validate the superiority of SAM-MI. Notably, the proposed method achieves a 16.7% relative improvement in mIoU over Grounded-SAM on the MESS benchmark, along with a 1.6$\times$ speedup. We hope SAM-MI can serve as an alternative methodology to effectively equip the OVSS model with SAM.