Improving convolutional neural networks performance for image classification using test time augmentation : a case study using MURA dataset

Kandel, Ibrahem (avtor), Castelli, Mauro (avtor)

Bone fractures are one of the main causes to visit the emergency room (ER); the primary method to detect bone fractures is using X-Ray images. X-Ray images require an experienced radiologist to classify them; however, an experienced radiologist is not always available in the ER. An accurate automatic X-Ray image classifier in the ER can help reduce error rates by providing an instant second opinion to the emergency doctor. Deep learning is an emerging trend in artificial intelligence, where an automatic classifier can be trained to classify musculoskeletal images. Image augmentations techniques have proven their usefulness in increasing the deep learning model's performance. Usually, in the image classification domain, the augmentation techniques are used during training the network and not during the testing phase. Test time augmentation (TTA) can increase the model prediction by providing, with a negligible computational cost, several transformations for the same image. In this paper, we investigated the effect of TTA on image classification performance on the MURA dataset. Nine different augmentation techniques were evaluated to determine their performance compared to predictions without TTA. Two ensemble techniques were assessed as well, the majority vote and the average vote. Based on our results, TTA increased classification performance significantly, especially for models with a low score.

• Jezik: Angleški jezik
• Ključne besede: image classification, convolutional neural networks, transfer learning, test time 29 augmentation, deep learning, ensemble learning
• Vrsta gradiva: Članek v reviji
• Tipologija: 1.01 - Izvirni znanstveni članek
• Organizacija: EF - Ekonomska fakulteta
• Status publikacije: Objavljeno
• Različica publikacije: Recenzirani rokopis
• Leto izida: 2021
• Št. strani: 22 str.
• Številčenje: Vol. 9, art 33
• PID: 20.500.12556/RUL-129209
• UDK: 659.2:004
• ISSN pri članku: 2047-2501
• DOI: 10.1007/s13755-021-00163-7
• COBISS.SI-ID: 72230659
• Datum objave v RUL: 30.08.2021

Gradivo je del revije

• Naslov: Health information science and systems
• Skrajšan naslov: Health inf. sci. syst.
• Založnik: Springer
• ISSN: 2047-2501
• COBISS.SI-ID: 523110681

Projekti

• Financer: FCT - Fundação para a Ciência e a Tecnologia, I.P.
• Številka projekta: DSAIPA/DS/0022/2018
• Naslov: GADgET

• Financer: ARRS - Agencija za raziskovalno dejavnost Republike Slovenije
• Program financ.: Raziskovalni program
• Številka projekta: P5-0410
• Naslov: Digitalizacija kot gonilo trajnostnega razvoja posameznika, organizacij in družbe