Semantic segmentation for fruit freshness identification using u-net

Fariq Fadhlan Aradhana and Mahmud Dwi Sulistiyo and Ema Rachmawati and Sugondo Hadiyoso and Hashim, Nik Mohd Zarifie and Murase, Hiroshi (2024) Semantic segmentation for fruit freshness identification using u-net. In: 2024 12th International Conference on Information and Communication Technology (ICoICT), 7 August 2024 through 8 August 2024, Bandung, Indonesia.

Text Semantic Segmentation for Fruit Freshness Identification Using U-Net.pdf Restricted to Registered users only Download (930kB)

Abstract

Indonesia has significant potential to develop a thriving fruit industry, yet the issue of fruit rot poses a major challenge, leading to substantial economic losses due to decreased fruit quality and the rapid spread of rot disease among other fruits. Immediate sanitization is essential to prevent this spread. This study explores a solution by employing semantic segmentation to accurately assess fruit freshness. Utilizing the U-Net model architecture, each pixel of the image is analyzed, incorporating an attribute-aware approach by creating specific attribute classes for different fruit characteristics. Specifically, two conditions of the fruits were: fresh and rotten. By assigning these attribute classes to the fruit images, the model is able to differentiate between these conditions during the segmentation process. Two training scenarios were examined: one using images of a single fruit, and the other using images containing one, two, and three fruits. The second scenario yielded superior results, with a mean Intersection over Union (mIoU) of 0.949, compared to 0.933 in the first scenario. These findings demonstrate the effectiveness of semantic segmentation in enhancing fruit freshness identification, offering a promising tool for mitigating fruit rot and improving the quality and safety of fruit products.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Fruit freshness, Identification, Semantic segmentation, U-Net
Divisions:	Faculty Of Electronics And Computer Technology And Engineering
Depositing User:	NUR FARISAH JAFRIN
Date Deposited:	23 Jul 2026 01:03
Last Modified:	23 Jul 2026 01:03
URI:	http://eprints.utem.edu.my/id/eprint/29898
Statistic Details:	View Download Statistic

Actions (login required)

View Item