Treffer: AA-TransDeeplabv3 + : a novel semantic segmentation framework for aerial images using adaptive and attentive based Transdeeplabv3 + with hybrid optimization technique.

Aerial imagesemantic segmentation is crucial for various operations such as military observation, land classification, and disaster impact assessments involving unmanned aerial vehicles. Although existing system is unsuited for aerial applications, these algorithms are mostly trained on human-centric datasets like "Cityscapes and Cam Vid". High-resolution aerial image semantic segmentation is a basic and difficult task with several applications. Even though numerous Convolution Neural Network (CNN) segmentation techniques have shown impressive results, it is still challenging to discriminate semantic parts among regions with comparable spectral properties employing only high-resolution data. Additionally, the typical data-independent up-sampling techniques could produce poor outcomes. Thus, a novel semantic segmentation technique is introduced to resolve the complication presented in the classical segmentation framework in aerial images by utilizing deep learning techniques. Here, an Adaptive and Attentive based TransDeeplabv3 + (AA-TransDeeplabv3 +)-based semantic segmentation model for input images is designed with a novel Hybridized Fire Hawk with Electric Fish Optimization (HFH-EFO). The parameters of Attentive-based TransDeeplabv3 + are tuned by developed HFH-EFO to attain the multi-objective function. The model is implemented using Python, which generates the empirical results. Therefore, the developed method achieves a dice coefficient of 93.02% and an accuracy value of 93.01%, outperforming traditionalapproaches. Hence, the proposedframework secures anexcellent result than the classical technique based on experimental analysis. [ABSTRACT FROM AUTHOR]

Copyright of Signal, Image & Video Processing is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)