Kim, Eunchan Kim, Seonghoon Choi, Myunghwan Seo, Taewon Yang, Sungwook 2024-01-19T10:31:07Z 2024-01-19T10:31:07Z 2023-01-26 2023-01 1424-8220 https://pubs.kist.re.kr/handle/201004/114150 We present a new deep learning framework for removing honeycomb artifacts yielded by optical path blocking of cladding layers in fiber bundle imaging. The proposed framework, HAR-CNN, provides an end-to-end mapping from a raw fiber bundle image to an artifact-free image via a convolution neural network (CNN). The synthesis of honeycomb patterns on ordinary images allows conveniently learning and validating the network without the enormous ground truth collection by extra hardware setups. As a result, HAR-CNN shows significant performance improvement in honeycomb pattern removal and also detailed preservation for the 1961 USAF chart sample, compared with other conventional methods. Finally, HAR-CNN is GPU-accelerated for real-time processing and enhanced image mosaicking performance. English Multidisciplinary Digital Publishing Institute (MDPI) Honeycomb Artifact Removal Using Convolutional Neural Network for Fiber Bundle Imaging Article 10.3390/s23010333 1 Sensors, v.23, no.1 Sensors 23 1 Y scie scopus 000909721600001 2-s2.0-85145966296 Chemistry, Analytical Engineering, Electrical & Electronic Instruments & Instrumentation Chemistry Engineering Instruments & Instrumentation Article ENHANCEMENT IMAGES fiber bundle imaging honeycomb artifact pattern synthesis convolution neural network (CNN)