Several recent successes in deep learning (DL), such as state-of-the-art performance on several image classification benchmarks, have been achieved through the improved configuration. Hyperparameters (HPs) tuning is a key factor affecting the performance of machine learning (ML) algorithms. Various state-of-the-art DL models use different HPs in different ways for classification tasks on different datasets. This manuscript provides a brief overview of learning parameters and configuration techniques to show the benefits of using a large-scale hand-drawn sketch dataset for classification problems. We analyzed the impact of different learning parameters and top-layer configurations with batch normalization (BN) and dropouts on the performance of the pre-trained visual geometry group 19 (VGG-19). The analyzed learning parameters include different learning rates and momentum values of two different optimizers, such as stochastic gradient descent (SGD) and Adam. Our analysis demonstrates that using the SGD optimizer and learning parameters, such as small learning rates with high values of momentum, along with both BN and dropouts in top layers, has a good impact on the sketch image classification accuracy.

Shaukat Hayat;Kun She;Muhammad Mateen;Parinya Suwansrikham;Muhammad Abdullah Ahmed Alghaili

School of Information and Software Engineering, University of Electronic Science and Technology of China, Chengdu 610054;School of Big Data & Software Engineering, Chongqing University, Chongqing 400044;College of Information Science and Engineering, Hunan University, Changsha 410082

电子科技学刊

[1]Shaukat Hayat;Kun She;Muhammad Mateen;Parinya Suwansrikham;Muhammad Abdullah Ahmed Alghaili-.Tuning-up Learning Parameters for Deep Convolutional Neural Network: A Case Study for Hand-Drawn Sketch Images)[J].电子科技学刊,2022(03):305-318

Drawn,Hyperparameters,HPs,dropouts

Tuning,Learning,Parameters,Deep,Convolutional,Neural,Network,Case,Study,Hand,Sketch,Images,Several,recent,successes,deep,learning,DL,such,state,art,performance,several,image,classification,benchmarks,have,been,achieved,through,improved,tuning,key,affecting,machine,ML,algorithms,Various,models,use,different,ways,tasks,datasets,This,manuscript,provides,brief,overview,techniques,show,benefits,using,large,scale,hand,drawn,sketch,problems,We,analyzed,impact,top,configurations,batch,normalization,BN,pre,trained,visual,geometry,group,VGG,include,momentum,values,optimizers,stochastic,gradient,descent,SGD,Adam,Our,analysis,demonstrates,that,small,high,along,both,layers,good,accuracy

0.549101