Video Instance Segmentation (VIS) is a new and inherently multi-task problem, which aims to detect, segment, and track each instance in a video sequence. Existing approaches are mainly based on single-frame features or single-scale features of multiple frames, where either temporal information or multi-scale information is ignored. To incorporate both temporal and scale information, we propose a Temporal Pyramid Routing (TPR) strategy to conditionally align and conduct pixel-level aggregation from a feature pyramid pair of two adjacent frames. Specifically, TPR contains two novel components, including Dynamic Aligned Cell Routing (DACR) and Cross Pyramid Routing (CPR), where DACR is designed for aligning and gating pyramid features across temporal dimension, while CPR transfers temporally aggregated features across scale dimension. Moreover, our approach is a light-weight and plug-and-play module and can be easily applied to existing instance segmentation methods. Extensive experiments on three datasets including YouTube-VIS (2019, 2021) and Cityscapes-VPS demonstrate the effectiveness and efficiency of the proposed approach on several state-of-the-art instance and panoptic segmentation methods. Codes will be publicly available at https://github.com/lxtGH/TemporalPyramidRouting.

Download full-text PDF

Source
http://dx.doi.org/10.1109/TPAMI.2022.3211612DOI Listing

Publication Analysis

Top Keywords

instance segmentation
12
pyramid routing
12
video instance
8
temporal pyramid
8
segmentation methods
8
instance
5
temporal
5
pyramid
5
improving video
4
segmentation
4

Similar Publications

With the rapid increase in end-of-life smartphones, enhancing the automation and intelligence of their recycling processes has become an urgent challenge. At present, the disassembly of discarded smartphones predominantly relies on manual labor, which is not only inefficient but also associated with environmental pollution and high labor intensity. In the context of end-of-life smartphone recycling, complex situations such as stacking and occlusion are commonly encountered.

View Article and Find Full Text PDF

MonoSeg: An Infrared UAV Perspective Vehicle Instance Segmentation Model with Strong Adaptability and Integrity.

Sensors (Basel)

January 2025

National Key Laboratory of Multispectral Information Intelligent Processing Technology, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, Wuhan 430000, China.

Despite rapid progress in UAV-based infrared vehicle detection, achieving reliable target recognition remains challenging due to dynamic viewpoint variations and platform instability. The inherent limitations of infrared imaging, particularly low contrast ratios and thermal crossover effects, significantly compromise detection accuracy. Moreover, the computational constraints of edge computing platforms pose a fundamental challenge in balancing real-time processing requirements with detection performance.

View Article and Find Full Text PDF

Focusing on Cracks with Instance Normalization Wavelet Layer.

Sensors (Basel)

December 2024

Shanxi Key Laboratory of Machine Vision and Virtual Reality, North University of China, Taiyuan 030051, China.

Automatic crack detection is challenging, owing to the complex and thin topologies, diversity, and background noises of cracks. Inspired by the wavelet theory, we present an instance normalization wavelet (INW) layer and embed the layer into the deep model for segmentation. The proposed layer employs prior knowledge in the wavelets to capture the crack features and filter the high-frequency noises simultaneously, accelerating the convergence of model training.

View Article and Find Full Text PDF

Two convex polyhedra that markedly resemble the head of the flatback sea turtle hatchling are identified. The first example is a zygomorphic tetragonal dodecahedron, while the other, an even better matching structure, is a related tetradecahedron, herein speculated to arise from this particular dodecahedron via known mechanisms gleaned from studies of the behavior of foams. A segmented, biomorphic, convex polyhedral model to address cephalic topology is thus presented stemming from solid geometry, anatomical observations, and a recently computed densest local packing arrangement of fifteen slightly oblate spheroids in which fourteen oblate spheroids surround a central such spheroid.

View Article and Find Full Text PDF

Autoimmune Tubulopathies.

J Am Soc Nephrol

January 2025

Centre de Recherche des Cordeliers, INSERM, Sorbonne Université, Université Paris Cité, F-75006 Paris, France.

The renal tubule and collecting duct express a large number of proteins, all having putative immunoreactive motives. Therefore, all can be the target of pathogenic autoantibodies. However, autoimmune tubulopathies seem to be rare and we hypothesize that they are underdiagnosed.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!