FindAPhD Weekly PhD Newsletter | JOIN NOW FindAPhD Weekly PhD Newsletter | JOIN NOW

Few shot learning for 3D plant phenomics


   Department of Computer Science

This project is no longer listed on FindAPhD.com and may not be available.

Click here to search FindAPhD.com for PhD studentship opportunities
  Prof B Tiddeman  No more applications being accepted  Funded PhD Project (Students Worldwide)

About the Project

The UKRI CDT in Artificial Intelligence, Machine Learning and Advanced Computing (AIMLAC) aims at forming the next generation of AI innovators across a broad range of STEMM disciplines. The CDT provides advanced multi-disciplinary training in an inclusive, caring and open environment that nurture each individual student to achieve their full potential. Applications are encouraged from candidates from a diverse background that can positively contribute to the future of our society. 

The UK Research and Innovation (UKRI) fully-funded scholarships cover the full cost of tuition fees, a UKRI standard stipend of £15,921 per annum and additional funding for training, research and conference expenses. The scholarships are open to UK and international candidates.

Closing date for applications is 12 February 2022. For further information on how to apply please click here and select the "UKRI CDT Scholarship in AIMLAC" tab.

Project Overview

The 3D analysis of plants has become increasingly effective in modelling the relative structure of organs and other traits of interest. Our previous work has developed a pipeline for 3D plant data capture (using multi-view images and novel feature point detection [1] and matching algorithms [2]), 3D segmentation using Deep Learning [3] (using our PatternNet algorithm [4]), and 3D measurement [5] (e.g. using RANSAC). A disadvantage of the current pipeline is the need for considerable amounts of training data for the deep learning segmentation and classification algorithms. Few-shot learning algorithms have been applied successfully in many computer vision problems to enable a classifier to generalize to new classes with very few examples. Recent work has developed a few-shot learning approach for 3D point clouds of indoor scenes / rooms. We have found 3D plant segmentation to be a much more challenging problem, with more complex branching structures and more noise, where previous supervised algorithms have failed. This PhD will investigate novel few-shot learning algorithms that are robust for this particularly challenging case.


References

M. Ghahremani, Y. Liu and B. Tiddeman, "FFD: Fast Feature Detector," in IEEE Transactions on Image Processing, vol. 30, pp. 1153-1168, 2021.
Morteza Ghahremani, Yitian Zhao, Bernard Tiddeman, Yonghuai Liu,
Interwoven texture-based description of interest points in images,
Pattern Recognition, Volume 113,2021.
Ghahremani Morteza, Williams Kevin, Corke Fiona M. K., Tiddeman Bernard, Liu Yonghuai, Doonan John H., "Deep Segmentation of Point Clouds of Wheat", Frontiers in Plant Science, Volume 12, 2021.
Ghahremani M., Tiddeman B., Liu Y., Behera A., "Orderly Disorder in Point Cloud Domain", ECCV, 2020.
Morteza Ghahremani, Kevin Williams, Fiona Corke, Bernard Tiddeman, Yonghuai Liu, Xiaofeng Wang, John H. Doonan, "Direct and accurate feature extraction from 3D point clouds of plants using RANSAC", Computers and Electronics in Agriculture, Volume 187,2021.
Na Zhao, Tat-Seng Chua, Gim Hee Lee, "Few-shot 3D Point Cloud Semantic Segmentation", CVPR, 2021.
Search Suggestions
Search suggestions

Based on your current searches we recommend the following search filters.

PhD saved successfully
View saved PhDs