Abstract:With the continuous development of intelligence, people's desire for the implementation of the monitoring and early warning function of drowning is becoming increasingly urgent for high-risk places such as rivers and other safety accidents. Panoptic segmentation, a fundamental technique in the field of computer vision, combines the functions of foreground detection and background semantic segmentation, allowing it to thoroughly assess the river scene and then match the algorithm requirements of intelligent early warning. However, the seasonality of the river scene, the diverse positions of foreground instances, and the haphazard distribution of background information will cause issues such as substantial changes in scene characteristics, multi-scale instance targets, and unclear boundary segmentation. To address these issues, we develop the spatial context prior module (SCPM), which increases the robustness of feature representation by emphasizing the spatial similarities and differences between similar and dissimilar pixels. In addition, a densely connected atrous spatial pyramid pooling (DenseASPP) is used to achieve multi-scale feature extraction. In the training stage, an edge feature fusion module (EFM) is proposed to fuse low-level edge features with high-level semantic information to make up for the lost edge information. Furthermore, we comply the seasonal river scene panoptic segmentation dataset (OUC-SRS-SEG), and test the proposed approaches on it. The results of experiments demonstrate the effectiveness of constructing the optimization methods. Our algorithm's PQ value is 3.28% and 3.67% greater than that of Panoptic-DeepLab and Panoptic SegFormer, respectively.

Predicting Future Instance Segmentation with Contextual Pyramid ConvLSTMs

Panoptic Segmentation for Seasonal River Scene Based on Spatial Context Prior and DenseASPP

Predicting Future Instance Segmentation by Forecasting Convolutional Features

APANet: Auto-Path Aggregation for Future Instance Segmentation Prediction

Predictive Feature Learning for Future Segmentation Prediction

Improving Video Instance Segmentation via Temporal Pyramid Routing

STC: Spatio-Temporal Contrastive Learning for Video Instance Segmentation.

CMS-LSTM: Context Embedding and Multi-Scale Spatiotemporal Expression LSTM for Predictive Learning

Solve the Puzzle of Instance Segmentation in Videos: A Weakly Supervised Framework With Spatio-Temporal Collaboration

Multi-feature and Multi-branch Action Segmentation Framework for Modeling Long-Short-Term Dependencies

Hybrid Instance-aware Temporal Fusion for Online Video Instance Segmentation

Dual Spatial-Temporal Feature Pyramid With Decoupled Temporal Mining for Video Semantic Segmentation

From Single to Multiple: Leveraging Multi-level Prediction Spaces for Video Forecasting

Video object segmentation by Multi-Scale Pyramidal Multi-Dimensional LSTM with generated depth context

Spatiotemporal BEV Pyramid Networks for Future Instance Prediction of Autonomous Driving

Multi-Timescale Context Encoding for Scene Parsing Prediction

MSSINet: Real-Time Segmentation Based on Multi-Scale Strip Integration

Future Semantic Segmentation with Convolutional LSTM

A Unified Efficient Pyramid Transformer for Semantic Segmentation

Following the Lecturer: Hierarchical Knowledge Concepts Prediction for Educational Videos

Next frame prediction using ConvLSTM