Leaf counting: Multiple scale regression and detection using deep CNNs

Yotam Itzhaky, Guy Farjon, Faina Khoroshevsky, Alon Shpigler, Aharon Bar Hillel

Research output: Contribution to conferencePaperpeer-review

9 Scopus citations

Abstract

Visual object counting is a computer vision task relevant to a broad spectrum of problems, and specifically to the phenotyping domain. We propose two novel deep learning approaches for the visual object counting task, demonstrating their efficiency on the CVPPP 2017 Leaf Counting Challenge dataset. The first method performs counting via direct regression, predicting the count value using multiple scale representations of the image and using a novel fusion technique to combine the multi-scale predictions. In the second method, we count after predicting and aggregating all the leaf center points. Experimental results show that both our algorithms outperform last year's CVPPP challenge winners, while our second pipe also provides additional information of the leaf center points with a 95% average precision.

Original languageEnglish
StatePublished - 1 Jan 2019
Event29th British Machine Vision Conference, BMVC 2018 - Newcastle, United Kingdom
Duration: 3 Sep 20186 Sep 2018

Conference

Conference29th British Machine Vision Conference, BMVC 2018
Country/TerritoryUnited Kingdom
CityNewcastle
Period3/09/186/09/18

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Fingerprint

Dive into the research topics of 'Leaf counting: Multiple scale regression and detection using deep CNNs'. Together they form a unique fingerprint.

Cite this