The decoder is a key component in deep networks for the semantic segmentation of street views. The existing methods rely on the limited receptive field for feature extraction without considering the contextual information, which could lead to errors in understanding complex scenes. Moreover, a balance of contextual information and computational cost must be considered to meet the needs of real-world applications. To address these problems, we introduce a Spatial Plaid Attention Decoder network, which uses a lightweight decoder with Spatial Plaid Attention to perform highly efficient operations for semantic segmentation. With approximately 4 million parameters (9.75% of the UPerNet), our decoder achieves state-of-the-art performance on public datasets such as Cityscapes and ADE20K, with 84.84% and 54.0% mIoU, respectively. In addition, our method reduces the total Flops by 34.95% and 32.85%, respectively. We demonstrate how contextual information helps the network in object recognition and how object features and contextual features contribute to the scene segmentation and recognition.

Striking a balance between efficiency and accuracy is a challenge in the design and implementation of decoders. Accurate decoders often tend to be highly complex and computationally costly. This paper presents a novel decoder for semantic segmentation: Spatial Plaid Attention Decoder (SPADe). We propose a Spatial Plaid Attention module that performs an efficient local feature collection through spatial feature folding to help the model capture the local structure of objects while using long-range feature aggregation to consider the global structures efficiently and accurately. This makes SPADe a suitable choice for applications with limited resources. With a size of 7.3% UPerNet decoder, our SPADe obtains state-of-the-art performance with several popular backbones on public benchmarks. On Cityscapes and ADE20K, SPADe obtains 84.3% and $53.8 \% \mathrm{mIoU}$, while reducing the total GFlops by 32.8% and 70.5%, respectively. We also demonstrate that the effective design of SPADe allows it to capture long-range dependencies with a large receptive field. Implementation of the SPADe is available at github/SPADe.

Contrastive learning (CL) is widely known to require many negative samples, 65536 in MoCo for instance, for which the performance of a dictionary-free framework is often inferior because the negative sample size (NSS) is limited by its mini-batch size (MBS). To decouple the NSS from the MBS, a dynamic dictionary has been adopted in a large volume of CL frameworks, among which arguably the most popular one is MoCo family. In essence, MoCo adopts a momentum-based queue dictionary, for which we perform a fine-grained analysis of its size and consistency. We point out that InfoNCE loss used in MoCo implicitly attract anchors to their corresponding positive sample with various strength of penalties and identify such inter-anchor hardness-awareness property as a major reason for the necessity of a large dictionary. Our findings motivate us to simplify MoCo v2 via the removal of its dictionary as well as momentum. Based on an InfoNCE with the proposed dual temperature, our simplified frameworks, Sim-MoCo and SimCo, outperform MoCo v2 by a visible margin. Moreover, our work bridges the gap between CL and non-CL frameworks, contributing to a more unified under-standing of these two mainstream frameworks in SSL. Code is available at: https://bit.ly/3LkQbaT.

While self-attention mechanism has shown promising results for many vision tasks, it only considers the current features at a time. We show that such a manner cannot take full advantage of the attention mechanism. In this paper, we present Deep Connected Attention Network (DCANet), a novel design that boosts attention modules in a CNN model without any modification of the internal structure. To achieve this, we interconnect adjacent attention blocks, making information flow among attention blocks possible. With DCANet, all attention blocks in a CNN model are trained jointly, which improves the ability of attention learning. Our DCANet is generic. It is not limited to a specific attention module or base network architecture. Experimental results on ImageNet and MS COCO benchmarks show that DCANet consistently outperforms the state-of-the-art attention modules with a minimal additional computational overhead in all test cases. The code is available at: https://github.com/13952522076/DCANet.

Urban regions are complicated functional systems that are closely associated with and reshaped by human activities. The propagation of online geographic information-sharing platforms and mobile devices equipped with the Global Positioning System (GPS) greatly proliferates proximate sensing images taken near or on the ground at a close distance to urban targets. Studies leveraging proximate sensing images have demonstrated great potential to address the need for local data in the urban land-use analysis. This paper reviews and summarizes the state-of-the-art methods and publicly available data sets from proximate sensing to support land-use analysis. We identify several research problems in the perspective of examples to support the training of models and means of integrating diverse data sets. Our discussions highlight the challenges, strategies, and opportunities faced by the existing methods using proximate sensing images in urban land-use studies.

In this paper, we present a cascaded context dependency module, which is a highly lightweight module that can improve the performance of deep convolutional neural networks for various visual tasks. Inspired by the feature pyramid work in object detection and the context dependency work in image recognition, we consider to cascade the contexts of multiscale feature maps to aggregate the locality and globality in a local region. We further extract the dependency between original input and cascaded contexts for feature recalibration. Without employing learnable layers, our method introduces almost no additional parameters and computations. Furthermore, Our module can be seamlessly plugged into many existing CNN architectures to improve the performance. Experiments on ImageNet and MS COCO benchmarks indicate that our method can achieve results on par with or better than related work. Qualitatively, we achieve an absolute 1.42% (77.3137% vs. 75.8974%) top-1 classification accuracy improvement based on ResNet50 on ImageNet 2012 validation set with negligible computational overhead. Besides, our method yields significant gains on the MS COCO benchmark for the object detection task. All codes and models are made publicly available1.1We submit all the codes, pre-trained models, and training log files to https://github.com/13952522076/ParameterFree.

The unrestricted open vocabulary and diverse substances of scenery images bring significant challenges to scene recognition. However, most deep learning architectures and attention methods are developed on general-purpose datasets and omit the characteristics of scene data. In this paper, we exploit the Attention Pyramid Module (APM) to tackle the predicament of scene recognition. Our method streamlines the multi-scale scene recognition pipeline, learns comprehensive scene features at various scales and locations, addresses the interdependency among scales, and further assists feature re-calibration as well as the aggregation process. APM is extremely light-weighted and can be plugged into existing network architectures in a parameter-efficient manner. By integrating APM into ResNet-50, we obtain a boost of top-1 accuracy by 3.54% on the benchmark dataset. Our comprehensive experiments demonstrate that APM achieves much improved performance comparing with the state-of-the-art attention methods using significantly less computation budget. Source code is provided on https://github.com/ZN-Qiao/APM.