Plant Segmentation Using Machine Learning Methods with Low Data Reliance

Li, Ruohua

Permalink

https://hdl.handle.net/2142/124943 Copy

Description

Title

Plant Segmentation Using Machine Learning Methods with Low Data Reliance

Author(s)

Li, Ruohua

Issue Date

2021-05-01

Keyword(s)

• Computer Vision
• Plant Phenotyping
• Image Segmentation
• Semantic Segmentation

Date of Ingest

2024-10-23T14:32:47-05:00

Abstract

This thesis examines the potential of some learning-based computer vision algorithms with low data reliance on the problem of plant segmentation, namely, few-shot learning algorithms and clustering algorithm. We thoroughly investigate the mechanisms, benchmarks, and features of each algorithm. Each algorithm is applied to the plant segmentation problem. Then we show and discuss the results of each algorithm and further apply possible modifications and tuning to them. After this study, we have gathered performances for 3 different algorithms which have recall values of 54.8%, 87.7%, and 96.9% respectively.

Type of Resource

text

Genre of Resource

dissertation/thesis

Language

eng

Owning Collections