Using Deep Learning to Determine Freshness of Fish
碩士 === 亞洲大學 === 生物資訊與醫學工程學系 === 107 === Unfresh fish may cause diseases that are harmful to human health. It is unfortunately difficult for ordinary people who are inexperienced to judge if a fish is fresh. This study uses a single multi-box detector (SSD) neural network to help determine the freshn...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2019
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/gygf9v |
| Summary: | 碩士 === 亞洲大學 === 生物資訊與醫學工程學系 === 107 === Unfresh fish may cause diseases that are harmful to human health. It is unfortunately difficult for ordinary people who are inexperienced to judge if a fish is fresh. This study uses a single multi-box detector (SSD) neural network to help determine the freshness of fish. We use 400 photos of fresh/unfresh fish and use the SSD algorithm for training in deep learning. Our experiments show that it can achieve 80.9mAP of accuracy.
|
|---|