A PCA-based frame selection method for applying CNN and LSTM to classify postural behaviour in sows.
Meiqing WangMaciek oczakMona Lilian Vestbjerg LarsenFlorian BayerKristina MaschatJohannes BaumgartnerJean-Loup RaultTomas NortonPublished in: Comput. Electron. Agric. (2021)
Keyphrases
- computational cost
- high precision
- classification method
- detection method
- high accuracy
- experimental evaluation
- similarity measure
- selection strategy
- synthetic data
- support vector machine svm
- cost function
- multiscale
- significant improvement
- high dimensional
- segmentation algorithm
- detection algorithm
- clustering method
- pairwise
- objective function
- bayesian networks
- selection algorithm
- image sequences
- fine tuning