News |  Sitemap |  Contact
PDF Export  | 

Anders Thorsen: Using ImageJ with the plugin ObjectJ for fish oocyte and egg measurements; Hough transform, particle analysis, manual size measurements, and categorization


Related to egg surveys and reproductive studies of fish it is often necessary to work up egg and oocyte samples. Samples of live pelagic eggs usually consist of perfectly round and highly transparent eggs which need to be diameter measured and categorized (species and/or stage). Sometimes these egg samples are fixed in formalin before they are analyzed. Then the eggs become more opaque, but with transparent parts. Oocyte samples from maturing fish mostly consist of opaque vitellogenic oocytes which can be more or less round or oval. Also the oocytes needs to be diameter measured and categorized.

To help us in analyzing these three sample types we used ImageJ with the plugin ObjectJ ( ObjectJ is a plugin for ImageJ that supports graphical vector objects that non-destructively mark images on a transparent layer. Composite objects can encapsulate different color-coded marker structures in order to bundle features that belong together. The results table supports statistics, sorting, color coding, qualifying and macro access.

For the live transparent round eggs we developed an automatic diameter measurement procedure based on Hough transform followed by a slide show facility with manual categorization of the eggs and possibility for diameter correction. For the display of categories a hand of a clock analogy was used.

When samples were fixed in formalin the eggs changed so that they had both transparent and opaque parts. This resulted in poor results using the Hough transform. For these fixed egg samples we therefore made a completely manual procedure, but with optimized tools for diameter measurements.

Samples consisting of opaque oocytes could be automatically diameter measured using traditional particle analysis based on grey tone and shape factor thresholding. Since some species of fish have oval rather than round oocytes we needed to introduce an ellipse matching factor. For cases were some oocytes omitted automatic detection a handy manual measurement tool was developed that worked both with round and oval oocytes. For categorization of the oocytes we used the same tools as already has been described for egg samples.

The software that was developed for these three types of analysis can be downloaded from the ObjectJ website.


ObjectJ, Hough, particle, egg, oocyte

Administrative data

Presenting author: Anders Thorsen
Organisation: Institute of Marine Research, Bergen, Norway,

co-authors: Norbert Vischer

University of Amsterdam, Amsterdam, Netherlands,

Cindy van Damme

IMARES, IJmuiden, Netherlands,

Type: Poster (portrait)

© Luxembourg Institute of Science and Technology | Legal Notice