LIME visualizations for the 3Conv3Dense model, providing insight into its prediction process.

LIME visualizations for the 3Conv3Dense model, providing insight into its prediction process.

<p>The highlighted regions indicate the key visual features that contributed most to the weight estimation. The visualizations show the model consistently focusing on anatomically relevant areas, such as the rib cage and abdomen.</p>

Saved in:

Bibliographic Details
Main Author:	Md Junayed Hossain (22615268) (author)
Other Authors:	Jannatul Ferdaus (22140270) (author), Ashraful Islam (1283403) (author), M. Ashraful Amin (22615271) (author)
Published:	2025
Subjects:	Sociology Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Information Systems not elsewhere classified precise feature extraction mean absolute error less accurate predictions error case analysis demonstrates superior accuracy input data quality effective farm management pretrained efficientnetb3 model model significantly surpasses traditional manual methods associated weight data livestock weight estimation explainable cnn framework traditional regression models traditional methods weight estimation image data linear regression framework designed effective approach weight prediction cattle weight yolov5 algorithm study emphasizes results show recent advances random forest provide insights productivity optimization one variation mse ), making process low mae including normalization histogram equalization health assessment first leveraged face challenges especially cnn deep learning custom cnn conventional models computer vision cnn models cnn ). cattle images based approach 85 kg 02 kg
Tags:	Add Tag No Tags, Be the first to tag this record!

Cannot write session to /tmp/vufind_sessions/sess_03pi137rq4669mdh18e5dfgq84