8 November 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
In this article, an integrated object detection and communication (IODC) system based on single-stage detection framework is first proposed for synthetic aperture radar (SAR) images in underwater environment. Specifically, the combination of the multiscale feature extraction module and semantic information enhance fusion module for SAR images in underwater environment is designed as the semantic encoder and the results prediction module with anchor-free detection method is designed as the semantic decoder. Considering the multiscale feature, the channel encoder composed by a multiscale fusion module and a redundancy module is designed, and the channel decoder is the inverse of the channel encoder. To adapt to the time-varying and complex wireless environment, an adaptive transmission (AT) mechanism based on attention mechanism and knowledge base is proposed for the IODC system. Moreover, considering the actual application requirements, a lightweight design for the IODC system with the AT mechanism is also conducted. The experiment results on the Sonar Common Target Detection dataset show that the proposed IODC system with AT mechanism can achieve nearly 91% average precision at 25 dB, which means the proposed system can achieve the effective integration of the objection detection and communication for SAR images in underwater environment. In the lightweight design, the model parameters of the proposed system can be reduced by up to 20%, with only a 2.22% sacrifice in performance.
19 August 2024 in IEEE Access
Common human pose estimation methods rely on 2D heatmap regression, which requires expensive upsampling layers to maintain the resolution of the heatmap and additional post-processing for coordinate decoding. These components hinder the inference speed of human pose estimation tasks. To address this challenge, we propose a new real-time human pose estimation framework, EdgePose. First, we design the convolutional module EdgeBlock-C and the edge attention module EdgeBlock-T, and then build a hybrid network based on them to take advantage of both ConvNet and VIT. In addition, EdgePose simplifies the human pose estimation process by converting the output of the key point coordinates into a pixel classification task along the horizontal and vertical axes, thereby eliminating the upsampling and post-processing operations that may hinder inference speed, and speeding up the model’s inference speed while ensuring accuracy. The experimental results show that EdgePose-S achieved an AP score of 68.6 in the COCO validation test, and at the same time achieved an inference speed of 285.7 FPS on an Intel i9-10920X CPU. In the embedded Jetson Xavier NX environment, EdgePose-B achieved an AP score of 72.2 and an inference speed of 51.3 FPS, which is better than existing two-stage pose estimation methods.