CAT-EDNet: Cross-Attention Transformer-Based Encoder-Decoder Network for Salient Defect Detection of Strip Steel Surface.
Qiwu LuoJiaojiao SuChunhua YangWeihua GuiOlli SilvénLi LiuPublished in: IEEE Trans. Instrum. Meas. (2022)
Keyphrases
- defect detection
- textured surfaces
- low complexity
- video codec
- distribution network
- distributed video coding
- decoding process
- feature extraction
- peer to peer
- automated visual inspection
- noisy channel
- error control
- surface defects
- video compression
- fuzzy logic
- visual attention
- fault diagnosis
- wyner ziv video coding
- motion estimation
- d objects