Learning A Deep Convolution Network with Turing Test Adversaries for Microscopy Image Super Resolution

Tom, Francis; Sharma, Himanshu; Mundhra, Dheeraj; Dastidar, Tathagato Rai; Sheet, Debdoot

doi:10.1109/isbi.2019.8759443

Cited by 10 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…$$. The objective is to discriminate between

{I}_{\mathrm{CT}}

and

{S}_{\mathrm{CT}}

when presented as a randomly shuffled matched pair and to predict the correct order shuffling in an approach similar to 39,40 . The objective is to minimize

$$ {J}_{T_I}(.)

…”

Section: Methodsmentioning

confidence: 99%

Simulating cross‐modal medical images using multi‐task adversarial learning of a deep convolutional neural network

Kumar,

Sharma,

Jehadeesan

et al. 2024

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Computed tomography (CT) and magnetic resonance imaging (MRI) are widely utilized modalities for primary clinical imaging, providing crucial anatomical and pathological information for diagnosis. CT measures X‐ray attenuation, while MRI captures hydrogen atom density in tissues. Despite their distinct imaging physics principles, the signals obtained from both modalities when imaging the same subject can be represented by modality‐specific parameters and common latent variables related to anatomy and pathology. This paper proposes an adversarial learning approach using deep convolutional neural networks to disentangle these factors. This disentanglement allows us to simulate one modality from the other. Experimental results demonstrate our ability to generate synthetic CT images from MRI inputs using the Gold‐atlas dataset, which consists of paired CT‐MRI volumes. Patch‐based learning techniques and a visual Turing test are employed to model discriminator losses. Our approach achieves a mean absolute error of 36.81 4.46 HU, peak signal to noise ratio of 26.12 0.31 dB, and structural similarity measure of 0.9 0.02. Notably, the synthetic CT images accurately represent bones, gaseous cavities, and soft tissue textures, which can be challenging to visualize in MRI. The proposed model operates at an inference compute cost of 430.68 GFlops/voxel. This method can minimize radiation exposure by reducing the need for pre‐operative CT scans, providing an MR‐only alternative in clinical settings.

show abstract

“…$$. The objective is to discriminate between

{I}_{\mathrm{CT}}

and

{S}_{\mathrm{CT}}

when presented as a randomly shuffled matched pair and to predict the correct order shuffling in an approach similar to 39,40 . The objective is to minimize

$$ {J}_{T_I}(.)

…”

Section: Methodsmentioning

confidence: 99%

Simulating cross‐modal medical images using multi‐task adversarial learning of a deep convolutional neural network

Kumar,

Sharma,

Jehadeesan

et al. 2024

Int J Imaging Syst Tech

View full text Add to dashboard Cite

show abstract

“…The findings reveal that the neural network outperforms conventional methods in estimating ambient light and transmission. Nevertheless, the CNN method's disadvantage is that it requires a large image dataset to process and achieves a successful output [24].…”

Section: Related Workmentioning

confidence: 99%

Automatic phytoplankton image smoothing through integrated dual image histogram specification and enhanced background removal method

Baharin

Ghani

Mohammad-Noor

et al. 2021

The Imaging Science Journal

View full text Add to dashboard Cite

Diatom is a dominant phytoplankton and commonly found in oceans or waterways. The captured phytoplankton microscopic images suffer from low contrast and surrounding debris. These images are not appropriated for identification. Integrated dual image contrast adaptive histogram specification with enhanced background removal (DIHS-BR) is proposed to address these issues by automatically removes the background of the phytoplankton image and improves the image quality while cropping phytoplankton cell. DIHS-BR will automatically remove the background and noises. DIHS-BR consists of two major steps, namely, contrast adaptive histogram specification and background removal by means of edge mask cropping. Results demonstrated that DIHS-BR filtered out the image background and left only the required phytoplankton cell image. Noises are minimized, while the contrast and colour of phytoplankton cells are improved. The average edge-based contrast measure (EBCM) of 83.065 demonstrates the best contrast improvement of the proposed methods compared with the other state-of-the-art methods.

show abstract

“…The conventional approaches include interpolation-based models [1][2][3][4][5][6] and reconstruction-based models [7][8][9]. With the development of deep learning, many scholars propose various convolution neural networks (CNNs) [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] for ISR.…”

Section: Introductionmentioning

confidence: 99%

Wavelet Sparse Coding-Based Lightweight Networks for Image Superresolution

Wang

2022

Mobile Information Systems

View full text Add to dashboard Cite

Image superresolution (ISR) is a hot topic. With the success of deep learning, the convolutional neural network-based ISR makes great progress recently. However, most state-of-the-art networks contain millions of parameters and hundreds of layers. It is difficult to apply models realistically. To solve this problem, we propose a Wavelet Sparse Coding-based Lightweight Network for Image Superresolution (WLSR). Our contributions include four aspects. Firstly, to improve the ISR performance, the WLSR utilizes the superiorities of wavelet sparse coding on ISR. Secondly, we take advantage of the dilated convolution to expand receptive fields. In this case, the filters in WLSR can acquire more information than common convolutional filters from input images. Thirdly, to deal with sparse code efficiently, we employ deformable convolution networks to obtain the convolutional kernels that concentrate on nonzero elements. Fourthly, to make WLSR uint8 quantization robust, we take advantage of the Clipped ReLU activation in the end of WLSR and balance the SR performance with running time. Experimental results indicate that, compared with state-of-the-art lightweight models, the WLSR can achieve exceptional performance with few parameters. Moreover, the WLSR contains 30 times fewer parameters than VDSR, but it works better than VDSR on the validation set.

show abstract

Learning A Deep Convolution Network with Turing Test Adversaries for Microscopy Image Super Resolution

Cited by 10 publications

References 10 publications

Simulating cross‐modal medical images using multi‐task adversarial learning of a deep convolutional neural network

Simulating cross‐modal medical images using multi‐task adversarial learning of a deep convolutional neural network

Automatic phytoplankton image smoothing through integrated dual image histogram specification and enhanced background removal method

Wavelet Sparse Coding-Based Lightweight Networks for Image Superresolution

Contact Info

Product

Resources

About