Embedded Files
Ongoing
Ongoing
Fourier-basis noise for data augmentation
Traditional image augmentations include but are not limited to rotation, crop, and flip, which mainly operate in the spatial domain. By using image data augmentation, the variety of the data increases, improving model robustness. Instead of operating in the spatial domain, this research aims at exploring whether augmenting the frequency information of images helps improve the robustness of models under adversarial attacks. To augment frequency information, Fourier basis noise is used [1], which is additive noise in a specific frequency. The frequency-augmented images are further used for training computer vision models, for which the robustness towards adversarial attack is evaluated and compared with state-of-the-art approaches.
Deblurring using Riemannian geometry
To predict how a plant will grow, it is important to investigate how division planes in plants form on a single cell level. To build prediction models of the plant cell division and their division planes, we need clear microscopy images. For real microscope setups, such clean images are usually not acquired, as the microscopy images always have a certain level of blur.
We know a blurred image is an in-focus image convolved with a blurring kernel. To get clean non-blurry images, one has to solve the deblurring deconvolution problem. The big issue is that one does not know the blurring kernel. To learn the blurring kernel and deblur images, it has been shown in the literature that Deep Learning is a successful tool.
The goal of this project is to create a deep learning method for deblurring that uses information from the geometry of the (image) data manifold1, 2 to improve deblurring. For the project, protoplast microscopy data will be provided on which the developed deblurring approach can be tested.
Deblurring using disentanglement learning
To predict how a plant will grow, it is important to investigate how division planes in plants form on a single cell level. To build prediction models of the plant cell division and their division planes, we need clear microscopy images. For real microscope setups, such clean images are usually not acquired, as the microscopy images always have a certain level of blur.
We know a blurred image is an in-focus image convolved with a blurring kernel. To get clean non-blurry images, one has to solve the deblurring deconvolution problem. The big issue is that one does not know the blurring kernel. To learn the blurring kernel and deblur images, it has been shown in the literature that Deep Learning is a successful tool.
Many deep learning techniques for deblurring consider the blurred image as a whole. They do not consider that a blurred image can be separated into an in-focus image and the blurring of this in-focus image. The goal of this project is to create a deep learning method for deblurring that uses this disentanglement to create a better deep learning based deblurring approach. For the project, protoplast microscopy data will be provided on which the developed deblurring approach can be tested.
Finished
Finished
Learning data-augmentation policies for computer vision using additive Fourier-basis noise
Defocus Blur Synthesis and Deblurring through Interpolation in the Latent Space
Transfer learning for cell image reconstruction
Convolutional Autoencoders models are designed and implemented to reconstruct cell slide images. The performance of the models to remove blur from sections of cell slides was investigated.
Cell Detection in Whole Slide Images With Out-of-Focus Corruption
Page updated
Google Sites
Report abuse