Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2023
- Goals
- Record
- 8f41849e-c89e-457d-b654-54d8da873ded
- Published
- 2026-10-30
Universitas Dian Nuswantoro conducted an activity on classifying milkfish freshness using deep learning and the ResNet50 architecture. The activity involved UDINUS researchers Al-Ghiffary, Maulana Malik Ibrahim; Sari, Christy Atika; Rachmawanto, Eko Hari; Cahyo, Nur Ryan Dwi; Yaacob, Noorayisahbe Mohd; and Ali, Rabei Raad, who participated in proposing a method for milkfish freshness-level classification.
The activity addressed milkfish, identified in the source as Chanos chanos. The source states that milkfish had become the main commodity in three major cities in Indonesia and contributed at least 77 thousand tons of aquaculture production in 2021. Fish quality was determined according to freshness, and freshness assessment was carried out during sorting. The sorting process generally evaluated the physical characteristics of fish. However, the source describes this approach as less efficient and economical because the ability to classify fish freshness could vary among individuals.
To address these stated problems, UDINUS proposed a classification method that utilized deep learning and the ResNet50 architecture. The method used a personal dataset divided into four different classes. According to the source, the proposed method produced an excellent result. The intended system was described as efficient and economical, with the aim that milkfish sorting could be performed more quickly and accurately.
This activity relates to the stated accreditation goal through its focus on milkfish, aquaculture production, fish quality assessment, and a classification method applied to the sorting process. Within the accreditation context, UDINUS presents this activity as institutional research concerning the application of a deep-learning method to milkfish freshness classification.