Banca de QUALIFICAÇÃO: REGINALDO AMARAL CAMPINA

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : REGINALDO AMARAL CAMPINA
DATE: 27/11/2026
TIME: 08:00
LOCAL: Salada Pós Graduação - EAJ/UFRN
TITLE:

IMAGING DURING IMBIBITION ALLOWS FOR EARLY PREDICTION OF THE VIABILITY AND VIGOR OF PURPLE IPÊ SEEDS FOR MACHINE LEARNING.


KEY WORDS:

Seed imbibition; seed vigor; Handroanthus impetiginosus; machine learning.


PAGES: 67
BIG AREA: Ciências Agrárias
AREA: Recursos Florestais e Engenharia Florestal
SUMMARY:

Accelerating seed analysis through rapid prediction of seed vigor is essential for native forest species such as Handroanthus impetiginosus (Mart. ex DC) Mattos, whose seeds rapidly lose physiological quality. The main objective of this study was to evaluate the potential of images captured after imbibition to predict seed vigor using machine learning, specifically identifying the period with the highest predictive capacity. Seeds underwent imbibition and were photographed at sequential intervals of 0, 3, 6, 9, 12, 15, 18, 21, and 24 hours (first set) and every 12 hours for seven days; subsequently, radicle protrusion was monitored, and seedling germination and vigor were assessed and classified. Morphological, chromatic, and textural characteristics were extracted from the seeds, and the resulting images were used to train a machine learning model (AI) for individual seed classification. Predictive performance was compared across different imbibition times to identify the optimal period for the early estimation of physiological quality. Images obtained during the initial phases of imbibition allowed for the prediction of viability and vigor prior to radicle protrusion, while also enabling a rapid estimate of the seed lot's physiological quality by aggregating individual predictions. The approach proposed in this study could reduce evaluation time from several days to a few hours, facilitating automated analysis applicable to forest seeds. 


COMMITTEE MEMBERS:
Externo ao Programa - 1204363 - ANTONINO ALVES FEITOSA NETO - nullInterna - 3081596 - ELAINE CRISTINA ALVES DA SILVA
Presidente - 1880265 - MARCIO DIAS PEREIRA
Notícia cadastrada em: 09/11/2026 14:48
SIGAA | Superintendência de Tecnologia da Informação - (84) 3342 2210 | Copyright © 2006-2026 - UFRN - sigaa08-producao.info.ufrn.br.sigaa08-producao