Real-time segmentation system and method based on hybrid expansion network
A network and convolution technology, applied in the field of computer vision, can solve problems such as difficulty in meeting segmentation requirements, insufficient information extraction capabilities, and decreased accuracy performance, and achieve the effects of improving accuracy, improving capture capabilities, and expanding receptive fields
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0022] Attached below figure 1 The preferred embodiment of the present invention is further described, and the present invention includes a lightweight backbone network MobileNet v2, a hybrid hollow convolution module;
[0023] The hybrid atrous convolution module consists of a lightweight spatial pyramid attention module and a global information enhancement module;
[0024] The described lightweight hybrid hole module achieves a comprehensive trade-off between accuracy and high efficiency (e.g., execution speed, memory footprint, or computational complexity) through multi-scale information and effective attention mechanism;
[0025] like figure 1 As shown, given an input, it is first fed into our backbone network to obtain semantic features. For high-resolution data sets, the output step size (OS) of the encoder is reasonably set to 8, which is a technique commonly used by people in the field to reduce the image size, so that the feature map can be down-sampled to 1 / 8 of th...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com