New radial processing software was developed, in order to improve the age estimation process of Cephalopod statoliths. A database was composed of 300 statolith images for Loligo forbesi, Loligo vulgaris, Sepia hierredda, and Sepia officinalis samples.

Various tests on statolith images showed smoothed profile integration with result the creation of a database with image mosaics.

Age-reading precision on known-age samples was high, especially for individuals of less than 200 days old. In addition, the distribution of statolith radius at 100 days showed a trend in the winter-hatched individuals with a higher variability in their growth, than spring-hatched individuals.

The image analysis tools defined in this program were developed, tested and are currently available near various users, before to be commercially distributed. Their development however underlined the need for complementary tools for image mosaic construction and a database, which were thus developed and which are available today. This tool set thus represents currently a coherent whole, enabling to assist the user since image acquisition until their most advanced treatment.

The choice was made firstly to produce the whole of the data processing sequences in order to lead to a complete and operational product and this with depends on the time devoted to its application to a real case study of the age and growth of a cephalopod. The application of these tools was however started on a comparative study of Loligo vulgaris growth patterns of specimens captured to different seasons along the Portuguese coasts. The sample will be supplemented in the next months and data processing will be continued in order to be published.

Among the reading problems identified, remains the slide preparation difficulties which are a strong limitation to statolith exploitation, particularly for conducting comparison between slides from different origin because of the lack of standardisation. It has been one of our main limitation in exploiting statolith slide collections available for the project.

Consequently, if the imaging tools does not solve all the cephalopod age estimation problems, curved and combined trajectories enable to collect information precisely where it is located and the interpolation enables to process samples where ring sequences are locally not visible, allowing for instance to back-calculate hatching dates and with this, to temporally align growth patterns.

Future work on computer-assisted cephalopod age estimation should be focused on the improvement of image quality of statolith images by filtering radial structures but also to the increasing of the standardisation of the preparation techniques.