Publication: Blister Package Classification Using ResNet-101 for Identification of Medication
Issued Date
2021-01-01
Resource Type
Other identifier(s)
2-s2.0-85125185670
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
ICSEC 2021 - 25th International Computer Science and Engineering Conference. (2021), 406-410
Suggested Citation
Narit Hnoohom, Nagorn Maitrichit, Pitchaya Chotivatunyu, Virach Sornlertlamvanich, Sakorn Mekruksavanich, Anuchit Jitpattanakul Blister Package Classification Using ResNet-101 for Identification of Medication. ICSEC 2021 - 25th International Computer Science and Engineering Conference. (2021), 406-410. doi:10.1109/ICSEC53205.2021.9684590 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/76707
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Blister Package Classification Using ResNet-101 for Identification of Medication
Abstract
This research aimed to fine-Tune image classification with deep learning techniques to verify the dispensing of prescriptions in hospitals. The proposed approach will be able to help pharmacies reduce the errors that lead to patients receiving the wrong medications. The image classification model uses a double-side transformed image dataset with download from Highlighted Deep Learning (HDL) paper. The dataset collected two-hundred seventy-Two images for types of medicine blister packs, including 72 images of front-side and backside merged with a horizontal cropped background, which were used for training the model. The blister package image dataset uses a deep learning model with a ResNet-101 pre-Trained model from the TensorFlow framework. The experimental results indicated that the TensorFlow framework achieved higher precision, recall, and F1-score than the Caffe framework. A ResNet-101 model with histogram equalization in the front and backside has the highest accuracy at 100 percent.