Showing posts with label plants. Show all posts
Showing posts with label plants. Show all posts

Friday, April 2, 2021

multiview multi-task on plant images

 

for herbarium and plant images, segmented leaves, barks, and flowers can be separated into views, which can then be fed into feature extraction layers such as shapes and vines, followed by typical neurla networks or graph networks, and multi-task prediction on family-genus-species

multi-task training here make sense because the predicted, family, genus species is hierarchical by nature. 

 https://en.wikipedia.org/wiki/Species

Q: ImageNet Classification is inherently multi-task, is it? 

vision transformers with attention, image with 16x16 words

https://arxiv.org/abs/2010.11929








Saturday, December 9, 2017

Quantifying and scaling global plant trait diversity

Quantifying and scaling global plant trait diversity


https://www.try-db.org/TryWeb/Home.php

From Alaska to Amazonia: First global maps of traits that drive vegetation growth
You are subscribed to News - News From the Field for National Science Foundation Update. This information has recently been updated, and is now available.

Friday, August 25, 2017

SERNEC examples and some ideas, batch download


http://sernecportal.org/portal/taxa/index.php?taxon=246202

http://sernecportal.org/portal/collections/individual/index.php?occid=6337203


some species disppaeared from current locations. dynamic distribution of species. visualization.

SERNEC offer batch-download of meta information that contains image_URL.

I pasted the URL into csv, and batch download the images with wget.

wget -i URL_list.csv

see http://xmodulo.com/how-to-download-multiple-files-with-wget.html