Comparative Study of Model Optimization Techniques in Fine-Tuned CNN Models

Poojary, Ramaprasad; Pai, Akul

doi:10.1109/icecta48151.2019.8959681

Cited by 64 publications

(41 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the ResNet-50 architectural model with an ideal combination in choosing the right number of epochs even though the computing time is running longer. This is in line with [15] that Adam optimization can improve accuracy because it has a bias correction technique that provides better approximations that can improve accuracy. The ResNet-50 architecture model using Adam optimization can improve the accuracy of the thorax image classification by 3.63% from the previous accuracy of 94.1% by using the standard ResNet-50 architecture model.…”

Section: Resultssupporting

confidence: 82%

Pneumonia Classification of Thorax Images using Convolutional Neural Networks

Suyuti

Setyati

2020

Inform

View full text Add to dashboard Cite

The digital image processing technique is a product of computing technology development. Medical image data processing based on a computer is a product of computing technology development that can help a doctor to diagnose and observe a patient. This study aimed to perform classification on the image of the thorax by using Convolutional Neural Network (CNN). The data used in this study is lung thorax images that have previously been diagnosed by a doctor with two classes, namely normal and pneumonia. The amount of data is 2.200, 1.760 for training, and 440 for testing. Three stages are used in image processing, namely scaling, gray scaling, and scratching. This study used Convolutional Neural Network (CNN) method with architecture ResNet-50. In the field of object recognition, CNN is the best method because it has the advantage of being able to find its features of the object image by conducting the convolution process during training. CNN has several models or architectures; one of them is ResNet-50 or Residual Network. The selection of ResNet-50 architecture in this study aimed to reduce the loss of gradients at certain network-level depths during training because the object is a chest image of X-Ray that has a high level of visual similarity between some pathology. Moreover, several visual factors also affect the image so that to produce good accuracy requires a certain level of depth on the CNN network. Optimization during training used Adaptive Momentum (Adam) because it had a bias correction technique that provided better approximations to improve accuracy. The results of this study indicated the thorax image classification with an accuracy of 97.73%.

show abstract

Section: Resultssupporting

confidence: 82%

Pneumonia Classification of Thorax Images using Convolutional Neural Networks

Suyuti

Setyati

2020

Inform

View full text Add to dashboard Cite

show abstract

“…VGG16 and Resnet50 models with Adam optimizer were used in [15] to fire detection from images and up to 92% accuracy was gained. More data should be obtained to improve accuracy and use different optimizers like SGD and RMSProp optimizers [16]. VGG16 and Resnet50 models were also used in [17] to detect facial emotions from images and up to 92.4% accuracy was gained.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Adam (Adaptive Moment Estimation): Adam optimizer is originated from the "Adaptive Moments" [16,21,22]. It is an update to the RMSProp optimizer.…”

Section: B Model Optimization Techniquesmentioning

confidence: 99%

“…This technique is an iterative method for optimizing. This optimizer is used broadly in machine learning and deep learning [16,22,23]. In this optimizer learning rate is a critical factor.…”

Section: Sgd (Stochastic Gradient Descent)mentioning

confidence: 99%

“…This technique is used to solve the problem of AdaGrad optimizer by using exponentially decaying averaging technique [16,22]. It maintains the balance between second-order moments of the gradients and the historical effects.…”

Section: Rmsprop (Root Mean Square Propagation)mentioning

confidence: 99%

See 2 more Smart Citations

Sports Recognition using Convolutional Neural Network with Optimization Techniques from Images and Live Streams

Reja¹,

Rahman²

2021

IJC

View full text Add to dashboard Cite

This paper deals with the issue of automated image and video recognition of sports. It is a category of appreciation of human behavior, which is a very difficult task in the present day to classify images and video clips into a categorized gallery. This research paper proposes a sports detection system using a deeper CNN model that combines a fully connected layer with fine-tuning. It is applied to classify five individual sports groups through images and videos. In this work, we use a video classification method based on the image. Extended Resnet50 and VGG16 two pre-trained Deep CNNs are applied to build this sports detection system. RMSProp, ADAM & SGD optimizers are used to train the extended CNN models for five Epochs on the proposed 5sports dataset by handpicking thousands of sports images from the internet to very smoothly classify the five different types of sports. Training accuracy of approximately 83% is observed for ResNet50 with an SGD optimizer for 5 sports classes and 95% is observed for 3 sports classes.

show abstract