anparashar's Projects
3D-Guided Face Manipulation of 2D Images for the Prediction of Post-Operative Outcome after Cranio-Maxillofacial Surgery
Attention-guided CNN for image denoising(Neural Networks,2020)
Salient attention U-Net model for tumor segmentation in breast ultrasound images, based on visual saliency maps
Medical Imaging, Denoising Autoencoder, Sparse Denoising Autoencoder (SDAE) End-to-end and Layer Wise Pretraining
Deep ConvNets based eye cancer detection
Papers and Public Datasets for Diabetic Retinopathy Detection
Image denoising using deep CNN with batch renormalization(Neural Networks,2020)
Deep feature selection using GA for classification of cervical cytology images.
A repository for computer vision applications - image classification, multi label classification, object detection and image segmentation
A project for denoising Medical Images with Guassian Noise.
Lipstick ain't enough: Beyond Color-Matching for In-the-Wild Makeup Transfer (CVPR 2021)
We proposed a novel U-Net-based model -- DC-UNet to do medical image segmentation.
Denoising medical images using AutoEncoders
DIAGNOSIS OF DIABETIC RETINOPATHY FROM FUNDUS IMAGES USING SVM, KNN, and attention-based CNN models with GradCam score for interpretability,
Image Denoising with Residual Learning
A tensorflow implementation of 'Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising' only for JPEG deblokcing
:octocat::octocat:A tensorflow implement of the paper "Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising"
Implementation of EfficientNet using PSO(Particle Swarm Optimization) instead of grid search
Proof of concept for control landmarks in diffusion models!
Real-time Facial Expression Transfer --> facial expression capture and reenactment via webcam
Multi-Label Classification of Eye Diseases
Pneumonia is known as a respiratory disease caused by bacterial, viral or fungal infection and has a high mortality rate. Identification of pneumonia is usually done by chest X-ray, but is hampered by other lung problems that the patient has experienced. Therefore, this study proposes the Convolutional Neural Networks method by adding a Gabor filter and several image contrast processing techniques such as image enhancement, mahotas and contrast limited adaptive histogram equalization. Tests using original data obtained an accuracy of 94.3% and 44.8% while the application of Image Enhancement obtained an accuracy of 92.95% and loss of 36.3%, preprocessing mahotas obtained an accuracy of 65% and a loss of 200%. preprocessing contrast limited adaptive histogram equalization obtained 88.3% accuracy and 79.6% loss. Original data is the best in terms of accuracy and image enhancement is the best in loss in this study.
GANimation: Anatomically-aware Facial Animation from a Single Image (ECCV'18 Oral) [PyTorch]
Source code for GARDNet: Robust Multi-View Network for Glaucoma Classification in Color Fundus Images