Influence of Batch Normalization on Convolutional Neural Networks in HRRP Target Recognition

Fu, Zhequan; Li, Shangsheng; Li, Xiangping; Dan, Bo; Wang, Xukun

doi:10.23919/aces48530.2019.9060588

Cited by 3 publications

(1 citation statement)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to ensure that the generated feature map size is 200×200, we add corresponding padding to each layer of the residual network. We further add batch normalization [32] and ReLU layer to each convolution layer of each unit except the last convolution layer. The structure and paraments of CanNet unit in Table Ⅰ, where CONV stands for convolution layer, BN stands for batch standardization, and ReLU stands for activation function.…”

Section: Refinement Based On Deep Residual Networkmentioning

confidence: 99%

Image Reconstruction Based on Fused Features and Perceptual Loss Encoder-Decoder Residual Network for Space Optical Remote Sensing Images Compressive Sensing

2021

View full text Add to dashboard Cite

Compressive sensing (CS) technology is introduced into space optical remote sensing image acquisition stage, which could make wireless image sensor network node quickly and accurately obtain images in the case of two constraints of limited battery power and expensive sensor costs. On this basis, in order to further improve the quality of CS image reconstruction, we propose fused features and perceptual loss encoder-decoder residual network (FFPL-EDRNet) for image reconstruction. FFPL-EDRNet consists of a convolution layer and a reconstruction network. We train FFPL-EDRNet end-to-end, thus greatly simplifying the pre-processing and post-processing process and eliminating the block effect of reconstructed images. The reconstruction network is based on residual network, which introduces multiscale feature extraction, multi-scale feature combination and multi-level feature combination. Feature fusion integrates low-level information with high-level information to reduce reconstruction error. The perceptual loss function based on pretrained InceptionV3 uses the weighted mean square error to define the loss value between the reconstructed image feature and the label image feature, which makes the reconstructed image more semantically similar to label image. In the measurement procedure, we use convolution to achieve block compression measurement, so as to obtain full image measurements. For image reconstruction, we firstly use a deconvolution layer to initially reconstruct the image and then use the residual network to refine the initial reconstructed image. The experimental results show that: in the case of measurement rates (MRs) of 0.25, 0.10, 0.04 and 0.01, the peak signal-to-noise ratio (PSNR) = 27.502, 26.804, 24.593, 21.359 and structural similarity (SSIM) = 0.842, 0.816, 0.720, 0.568 of the reconstructed images obtained by FFPL-EDRNet. Therefore, Our FFPL-EDRNet could enhance the quality of image reconstruction.

show abstract

Section: Refinement Based On Deep Residual Networkmentioning

confidence: 99%

Image Reconstruction Based on Fused Features and Perceptual Loss Encoder-Decoder Residual Network for Space Optical Remote Sensing Images Compressive Sensing

2021

View full text Add to dashboard Cite

show abstract

Electrically Large Complex Objects Recognition Using Deep Learning Approach

Liu,

Hao,

Gong

et al. 2023

2023 IEEE 7th International Symposium on Electromagnetic Compatibility (ISEMC)

View full text Add to dashboard Cite

Robustness Fine-Tuning Deep Learning Model for Cancers Diagnosis Based on Histopathology Image Analysis

2023

View full text Add to dashboard Cite

Histopathology is the most accurate way to diagnose cancer and identify prognostic and therapeutic targets. The likelihood of survival is significantly increased by early cancer detection. With deep networks’ enormous success, significant attempts have been made to analyze cancer disorders, particularly colon and lung cancers. In order to do this, this paper examines how well deep networks can diagnose various cancers using histopathology image processing. This work intends to increase the performance of deep learning architecture in processing histopathology images by constructing a novel fine-tuning deep network for colon and lung cancers. Such adjustments are performed using regularization, batch normalization, and hyperparameters optimization. The suggested fine-tuned model was evaluated using the LC2500 dataset. Our proposed model’s average precision, recall, F1-score, specificity, and accuracy were 99.84%, 99.85%, 99.84%, 99.96%, and 99.94%, respectively. The experimental findings reveal that the suggested fine-tuned learning model based on the pre-trained ResNet101 network achieves higher results against recent state-of-the-art approaches and other current powerful CNN models.

show abstract

Influence of Batch Normalization on Convolutional Neural Networks in HRRP Target Recognition

Cited by 3 publications

References 4 publications

Image Reconstruction Based on Fused Features and Perceptual Loss Encoder-Decoder Residual Network for Space Optical Remote Sensing Images Compressive Sensing

Image Reconstruction Based on Fused Features and Perceptual Loss Encoder-Decoder Residual Network for Space Optical Remote Sensing Images Compressive Sensing

Electrically Large Complex Objects Recognition Using Deep Learning Approach

Robustness Fine-Tuning Deep Learning Model for Cancers Diagnosis Based on Histopathology Image Analysis

Contact Info

Product

Resources

About