Αναγνώριση και παρακολούθηση αντικειμένων σε βίντεο μικροσκοπίας
Object recognition and tracking in microscopy videos
View/
Keywords
Tracking ; Bioimaging ; Image analysis ; VideoAbstract
Automated detection and tracking of objects in sequences of images is an application of Image Analysis that has witnessed a significant rise in interest and has been the subject of extenssive research in the past years. Indicative of this is the fact that searches for the terms image tracking on pubmed returns a total of 11,404 publications for the years 1959-2010 and 36,693 publications from 2011 until the date of writing.
As we will also see later, there is a vast amount of methods and algorithms for this process, an understandable fact given the fact that object detection and tracking is being applied in numerous different fields like astronomy, biology, medicine, robotics, etc. Most of these fields however handle entirely different types of images and care about completely different types of objects, creating the need for specialized methods.
The large number of methodologies stemming from this, combined with the fact that the people that need to use them are rarely specialized in image analysis and programming, makes the application of most methodologies inaccessible and impractical.
When it comes to Biology in particular, which is the focus of this work, the techno- logical progress around the means of obtaining images, as well as the shift in the focus of current research, from pure genetics to the combined study of the entire environment inside living organisms has attenuated the need for accessible means of object tracking in biological images.
The purpose of this work is to provide a brief description of the most commonly used or state-of-the art methods for processing, analyzing and tracking objects in videos, with a focus on those applied to images and videos of biological origin. In the second part we present a computational framework intended to simplify and streamline the application of such methods in an attempt to make them more accessible to people without a backgroung in image analysis or software engineering.