An improved categorical cross entropy for remote sensing image classification based on noisy labels.
Panle LiXiaohui HeXijie ChengMengjia QiaoDingjun SongMingyang ChenTao ZhouJiamian LiXiaoyu GuoShaokai HuZhihui TianPublished in: Expert Syst. Appl. (2022)
Keyphrases
- remote sensing
- cross entropy
- image classification
- remotely sensed data
- change detection
- multispectral
- log likelihood
- remote sensing images
- high resolution
- maximum likelihood
- remote sensing data
- image analysis
- hyperspectral
- image processing
- satellite images
- remote sensing imagery
- feature extraction
- satellite data
- image features
- remotely sensed
- language modeling
- hyperspectral images
- land cover
- automatic image registration
- evaluation metrics
- training data
- remotely sensed images
- remote sensed images
- multi spectral images
- high spatial resolution
- error function
- language model
- digital image analysis
- query processing
- training set
- earth science
- density estimation
- neural network