Unverified50% confidenceFactExact time
Early CNN-based food image classification solutions achieved accuracy of only approximately 70–80%.
1
Sources
50%
Confidence
Long-term
Relevance
8/10/2026
First Seen
Sources
Macrobite：拍照即算营养，让宏量素追踪不再繁琐
Related Claims
UnverifiedThe portion estimation accuracy of photo-based nutrition recognition typically has a margin of error of 20–30%.64% similarUnverifiedExperimental data shows that when key information is located at the beginning and end of the input sequence, the model's retrieval accuracy can be as high as over 90%, but when located in the middle, it may drop sharply to below 60%.61% similarUnverified仅560个基础模型就承担了所有分析图像中80%的生成量58% similarUnverifiedIn a well-calibrated model, an output score of 0.8 should statistically correspond to an actual accuracy of 80%.58% similarUnverified常见的数据集划分比例为70/15/15或80/10/1057% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/725972API
curl https://kongchang.com/api/v1/knowledge/claims/725972MCP
get_claim(id=725972)