In the conventional process, the grain quality evaluation team of the Rice Varietal Improvement Group (RVIG) manually evaluates the physical attributes of 600-800 promising lines every year. They evaluate two sets of 30g milled rice of candidate elite line using their naked eye, a process that is tedious and time consuming for researchers.
“This prompted us to come up with the PhilRice Milled Grain Classifier (PMGC), a software that can speed up the conventional classification process,” said Imeldalyn G. Pacada, PhilRice senior science research specialist.
A classifier evaluates 30g of milled rice and can assess its physical attributes at around 48-96min. By using PMGC, a classifier can evaluate 6.2g of milled rice in less than 5min.
The software provides quick overview of analyzed milled grain samples that can be enlarged for verification. It validates translucent, chalky, and immature grains and gives grain ID number and color. It can also determine grain length and shape, and identify broken and brewer grains.
According to Pacada, PMGC was developed by establishing an algorithm using special programming language for image acquisition, processing, and integration of Artificial Neural Network (ANN). The developed algorithm includes the development of Graphical User Interface (GUI) to control the hardware and execute the image analysis software. The establishment of models or training samples was the key for increasing the predicting value of the software.
“This consists of image acquisition of different degree of chalky grains and various samples of immature grains that were used for model development with the help of neuroshell program,” Pacada explained.
The research team composed of Pacada, Evelyn H. Bandonill, Thessa Marie M. Pascual, Fred Jan A. Fracia, Arvin Paul P. Tuaño, Andres M. Tuates, and Thelma F. Padolina hopes that the software can help classifiers and plant breeders for faster grain quality evaluation.
The software was developed under the research study titled New tools for predicting chalkiness and immature grains in milled rice. The study won the best poster award during the 29th National Rice R&D Conference held at PhilRice, Sept. 7-8.