Abstract
In this study, we propose a multi-scale ensemble learning method for thermal image enhancement in different image scale conditions based on convolutional neural networks. Incorporating the multiple scales of thermal images has been a tricky task so that methods have been individually trained and evaluated for each scale. However, this leads to the limitation that a network properly operates on a specific scale. To address this issue, a novel parallel architecture leveraging the confidence maps of multiple scales have been introduced to train a network that operates well in varying scale conditions. The experimental results show that our proposed method outperforms the conventional thermal image enhancement methods. The evaluation is presented both quantitatively and qualitatively.
Highlights
Many kinds of research have been conducted on how to obtain a thermal image of high quality, which is needed in a wide range of applications: face detection and tracking [1], breast abnormality evaluation [2], pipeline leak recognition [3], and advanced driver assistance systems [4,5,6]
We propose a multi-scale ensemble learning method for thermal image enhancement in varying scale conditions
Thermal image enhancement can be largely categorized into three topics: detail enhancement, noise reduction, and contrast enhancement
Summary
Many kinds of research have been conducted on how to obtain a thermal image of high quality, which is needed in a wide range of applications: face detection and tracking [1], breast abnormality evaluation [2], pipeline leak recognition [3], and advanced driver assistance systems [4,5,6]. Lee et al [11] presented a residual network for thermal image enhancement They experimentally verified that the brightness domain is best suited for training a network for thermal image enhancement. This method generates a high-quality thermal image by element-wise summation of low-quality input and residual output images. Gupta and Mitra [13] proposed a hierarchical edge-based guided super-resolution method This method needs visiblerange images to extract multi-level edge information. Cascarano et al [15] presented a super-resolution algorithm for aerial and terrestrial thermal images, which was based on total variation regularization This method is a fully automatic method with a low-cost adaptive rule. They introduced a new thermal image quality metric based on a specific region of interest for radiometric analysis
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
