T
Tivadar Danka
Researcher at Hungarian Academy of Sciences
Publications - 8
Citations - 350
Tivadar Danka is an academic researcher from Hungarian Academy of Sciences. The author has contributed to research in topics: Segmentation & Modular design. The author has an hindex of 5, co-authored 7 publications receiving 204 citations.
Papers
More filters
Journal ArticleDOI
nucleAIzer: A Parameter-free Deep Learning Framework for Nucleus Segmentation Using Image Style Transfer.
Reka Hollandi,Abel Szkalisity,Tímea Tóth,Tímea Tóth,Ervin Tasnadi,Ervin Tasnadi,Csaba Molnar,Csaba Molnar,Botond Mathe,Istvan Grexa,Istvan Grexa,Jozsef Molnar,Arpad Balind,Mate Gorbe,Maria Kovacs,Ede Migh,Allen Goodman,Tamas Balassa,Tamas Balassa,Krisztian Koos,Wenyu Wang,Juan C. Caicedo,Norbert Bara,Ferenc Kovács,Lassi Paavolainen,Tivadar Danka,Andras Kriston,Anne E. Carpenter,Kevin Smith,Kevin Smith,Peter Horvath,Peter Horvath +31 more
TL;DR: The key to the approach is that during training nucleAIzer automatically adapts its nucleus-style model to unseen and unlabeled data using image style transfer to automatically generate augmented training samples, making deep learning for nucleus segmentation fairly simple and labor free for most biological light microscopy experiments.
Journal ArticleDOI
Phenotypic Image Analysis Software Tools for Exploring and Understanding Big Image Data from Cell-Based Assays
Kevin Smith,Filippo Piccinini,Tamas Balassa,Krisztian Koos,Tivadar Danka,Hossein Azizpour,Peter Horvath,Peter Horvath +7 more
TL;DR: The strengths and weaknesses of non-commercial phenotypic image analysis software are examined, recent developments in the field are covered, challenges are identified, and a perspective on future possibilities are given.
Posted Content
modAL: A modular active learning framework for Python.
Tivadar Danka,Peter Horvath +1 more
TL;DR: modAL as mentioned in this paper is a modular active learning framework for Python, aimed to make active learning research and practice simpler and make fast prototyping and easy extensibility possible, aiding the development of real-life active learning pipelines and novel algorithms.
Posted ContentDOI
A deep learning framework for nucleus segmentation using image style transfer
Reka Hollandi,Abel Szkalisity,Tímea Tóth,Ervin Tasnadi,Csaba Molnar,Botond Mathe,Istvan Grexa,Jozsef Molnar,Arpad Balind,Mate Gorbe,Maria Kovacs,Ede Migh,Allen Goodman,Tamas Balassa,Krisztian Koos,Wenyu Wang,Norbert Bara,Ferenc Kovács,Lassi Paavolainen,Tivadar Danka,Andras Kriston,Anne E. Carpenter,Kevin Smith,Kevin Smith,Peter Horvath,Peter Horvath +25 more
TL;DR: This work presents a deep learning approach aiming towards a truly general method for localizing nuclei across a diverse range of assays and light microscopy modalities and outperforms the 739 methods submitted to the 2018 Data Science Bowl on images representing a variety of realistic conditions.
Journal ArticleDOI
A Deep Learning-Based Approach for High-Throughput Hypocotyl Phenotyping.
TL;DR: A deep learning-based algorithm provides an adaptable tool for determining hypocotyl or coleoptile length of different plant species, and it is shown that the accuracy of the method reaches human performance.