深度學習影像辨識之電子模組租借管理系統

Electronic Components Renting System by Machine Learning Image Recognition

吳亦超(Yi-Chao Wu) ； 朱品澔(Pin-Hao Zhu) ； 許正裕(Cheng-Yu Hsu)

《修平學報》 44期 (2022/03) Pp. 1-22

摘要

隨著QR Code技術的崛起，目前實驗室裡大多數電子模組租借管理多採用QR Code掃瞄方式。但隨著微機電技術的進步，許多嵌入式系統的相關電子模組逐漸縮小化，導致無法在這類的電子模組上貼附QR Code標籤，進而無法透過QR Code掃瞄方式來完成租借管理。有鑑於此，本論文提出「深度學習影像辨識之電子模組租借管理系統（RSMLIC）」來解決上述問題。RSMLIC主要整合雲端Azure及資料庫，透過ML.NET中之ONNX框架之預先定型的訓練模型，透過訓料資料集的上傳與訓練，讓不同電子模組只需要透過影像拍攝即可完成影像辨識，並完成租借管理。透過實驗結果顯示，RSMLIC之電子模組正確辨識率達到95%以上，並可在3秒內完成辨識，因此證明RSMLIC將可解決上述無法貼附QR Code標籤之較小電子模組，進而無法透過QR Code掃瞄方式來完成租借管理的問題。

關鍵字

深度學習 ； 影像辨識 ； 資料庫 ； ONNX ； ML.NET

並列摘要

Since the smaller embedded electronic components could not be rented by QR code scanning system since it is not prone to be stick the QR code tag with these smaller components. An electronic components renting system by machine learning image recognition, RSMLIC, was thus proposed in this paper. In RSMLIC, the training data were uploaded to ONNX model in ML.NET. The deep learning was then executed in the cloud Azure web site. By using ONNX model with ML.NET and cloud database, the different electronic components could be distinguished to be rented by image recognition. The experimental results showed that the rate of correct image recognition could be up to 95% within 3 seconds. It proved that RSMLIC could be substituted for the renting system by QR code scanning.

並列關鍵字

Deep Learning ； Image Recognition ； Database ； ONNX ； ML.NET

參考文獻

Babukarthik, R. G., Adiga, V. A. K., Sambasivam, G. Chandramohan D., and Amudhavel J. （2020）. Prediction of COVID-19 Using Genetic Deep Learning Convolutional Neural Network. IEEE Access, 8, 177647-177666. doi： 10.1109/ACCESS.2020.3025164.

Moura, J. D., García, L. R., Vidal, P. F., Cruz, M., López, L. A., Lopez, E. C., Novo, J., and Ortega, M. （2020）. Deep Convolutional Approaches for the Analysis of COVID-19 Using Chest X-Ray Images From Portable Devices. IEEE Access, 8, 195594-195607, doi：10.1109/ACCESS.2020.3033762.

Xing, W. and Du, D. （2018）. Dropout Prediction in MOOCs： Using Deep Learning for Personalized Intervention. Journal of Educational Computing Research, 57（3）, 547- 570, doi： org/10.1177/0735633118757015.

Yamashita, R., Nishio, M., Do, R. K. G., and Togashi, K. （2018）. Convolutional Neural Networks： An Overview and Application in Radiology. Insights Imaging, 9（4）, 611- 629, doi： 10.1007/s13244-018-0639-9.

Luis Quintanilla（2021）。教學課程：在 ML.NET 中使用 ONNX 偵測物件【新聞群組或線上論壇】。https：//docs.microsoft.com/zh-tw/dotnet/machine-learning/tutorials/object-detection-onnx。