A5059079555- Advanced Image Processing Techniques
- Image Processing Techniques and Applications
- Image and Signal Denoising Methods
- Advanced Vision and Imaging
- Advanced Image Fusion Techniques
- Video Surveillance and Tracking Methods
- Semiconductor Quantum Structures and Devices
- Advanced Image and Video Retrieval Techniques
- Optical Coherence Tomography Applications
- Retinal Imaging and Analysis
- Sparse and Compressive Sensing Techniques
- Cell Image Analysis Techniques
- Image and Object Detection Techniques
- Image Retrieval and Classification Techniques
- AI in cancer detection
- Image Enhancement Techniques
- Cellular transport and secretion
- Photoacoustic and Ultrasonic Imaging
- Semiconductor materials and interfaces
- Video Analysis and Summarization
- Infrared Target Detection Methodologies
- Digital Radiography and Breast Imaging
- Remote-Sensing Image Classification
- Medical Image Segmentation Techniques
- Image Processing and 3D Reconstruction
Czech Academy of Sciences, Institute of Information Theory and Automation
2015-2024
Czech Academy of Sciences
2009-2022
Czech Academy of Sciences, Institute of Photonics and Electronics
2003-2013
Instituto de Óptica "Daza de Valdés"
2007
Czech Technical University in Prague
1998
In this work we address the problem of blind deconvolution and denoising. We focus on restoration text documents show that type highly structured data can be successfully restored by a convolutional neural network. The networks are trained to reconstruct high-quality images directly from blurry inputs without assuming any specific blur noise models. demonstrate performance large set combination realistic de-focus camera shake kernels. On artificial data, significantly outperform existing...
Blind deconvolution, which comprises simultaneous blur and image estimations, is a strongly ill-posed problem. It by now well known that if multiple images of the same scene are acquired, this multichannel (MC) blind deconvolution problem better posed allows estimation directly from degraded images. We improve MC idea adding robustness to noise stability in case large blurs or size vastly overestimated. formulate as an l(1) -regularized optimization seek solution alternately optimizing with...
Blind image deconvolution is required in many applications of microscopy imaging, remote sensing, and astronomical imaging. Unfortunately a single-channel framework, serious conceptual numerical problems are often encountered. Very recently, an eigenvector-based method (EVAM) was proposed for multichannel framework which determines perfectly convolution masks noise-free environment if channel disparity, called co-primeness, satisfied. We propose novel iterative algorithm based on recent...
Existing multichannel blind restoration techniques assume perfect spatial alignment of channels, correct estimation blur size, and are prone to noise. We developed an alternating minimization scheme based on a maximum posteriori with priori distribution blurs derived from the framework original images defined by variational integral. This stochastic approach enables us recover image channels severely corrupted observe that exact knowledge size is not necessary, we prove translation...
This paper presents a new approach to the blind deconvolution and superresolution problem of multiple degraded low-resolution frames original scene. We do not assume any prior information about shape degradation blurs. The proposed consists building regularized energy function minimizing it with respect image blurs, where regularization is carried out in both blur domains. based on variational principles maintains stable performance under severe noise corruption. guarantees consistency...
Highlights•AP2 μ2T156 phosphorylation starts early at CCP inception and increases as grows•Phosphorylation favors an AP2 conformational change also triggers NECAP binding•NECAP PHear simultaneously binds P-AP2 membrane-remodeling proteins on opposite faces•Disrupting the P-AP2:NECAP:membrane-remodeling protein network reduces CME ratesSummaryClathrin-mediated endocytosis (CME) is key to maintaining transmembrane composition of cells' limiting membranes. During mammalian CME, a reversible...
Clathrin-mediated endocytosis (CME) is the main mechanism by which mammalian cells control their cell surface proteome. Proper operation of pivotal CME cargo adaptor AP2 requires membrane-localized Fer/Cip4 homology domain-only proteins (FCHO). Here, live-cell enhanced total internal reflection fluorescence–structured illumination microscopy shows that FCHO marks sites clathrin-coated pit (CCP) initiation, mature into uniform-sized CCPs comprising a central patch and clathrin corralled an...
Long-exposure handheld photography is degraded with blur, which difficult to remove without prior information about the camera motion. In this work, we utilize inertial sensors (accelerometers and gyroscopes) in modern smartphones detect exact motion trajectory of smartphone during exposure blur from resulting based on recorded data. The whole system implemented Android platform embedded device, a close-to-real-time deblurring algorithm. performance proposed demonstrated real-life scenarios.
Retinal imaging plays a key role in the diagnosis and management of ophthalmologic disorders, such as diabetic retinopathy, glaucoma, age-related macular degeneration. Because acquisition process, retinal images often suffer from blurring uneven illumination. This problem may seriously affect disease progression assessment. Here we present method for color image restoration by means multichannel blind deconvolution. The is applied to pair acquired within lapse time, ranging several minutes...
Blur is an image degradation that difficult to remove. Invariants with respect blur offer alternative way of a~description and recognition blurred images without any deblurring. In this paper, we present original unified theory invariants. Unlike all previous attempts, the new does not require prior knowledge type. The invariants are constructed in Fourier domain by means orthogonal projection operators moment expansion used for efficient stable computation. It shown published earlier just...
In many real applications, blur in input low-resolution images is a nuisance, which prevents traditional super-resolution methods from working correctly. This paper presents unifying approach to the blind deconvolution and superresolution problem of multiple degraded frames original scene. We introduce method assumes no prior information about shape degradation blurs properly defined for any rational (fractional) resolution factor. The minimizes regularized energy function with respect...
We propose a practical method to remove photo blur due camera shake, which is typical problem when taking photos in dim lighting conditions such as indoor or night scenes. use pair of images, one them blurred and the other underexposed noisy because high ISO setting. Existing methods assume convolution model, that same whole image. It seldom true practice, especially for wide angle lens photos. apply space-variant model blurring valid many real situations. Results are documented by photograph scene.
OBJECTIVE The highly infiltrative growth of glioblastoma (GBM) makes distinction between the tumor and normal brain tissue challenging. Therefore, fluorescence-guided surgery is often used to improve visual identification radiological margins. aim this study was evaluate ability recently developed molecularly targeted near-infrared (NIR) protease-activated probes visualize GBM compare most promising candidate with gold standard, 5-aminolevulinic acid (5-ALA). METHODS Single-substrate 6QC-ICG...
The notion of a Fast Moving Object (FMO), i.e. an object that moves over distance exceeding its size within the exposure time, is introduced. FMOs may, and typically do, rotate with high angular speed. are very common in sports videos, but not rare elsewhere. In single frame, such objects often barely visible appear as semi-transparent streaks. A method for detection tracking proposed. consists three distinct algorithms, which form efficient localization pipeline operates successfully broad...
Standard convolution as a model of radiometric degradation is in majority cases inaccurate the blur varies space and we are thus required to work with computationally demanding space-variant model. Space-variant can be approximately decomposed set standard convolutions. We explain detail properties operator show two possible decomposition models approximation approaches. Our target application image deconvolution, on which illustrate theoretical differences between these models. propose...