DCT-ViT: High-Frequency Pruned Vision Transformer With Discrete Cosine Transform.
Jongho LeeHyun KimPublished in: IEEE Access (2024)
Keyphrases
- high frequency
- discrete cosine transform
- image compression
- subband
- low frequency
- filter bank
- dct coefficients
- blocking artifacts
- wavelet transform
- transform domain
- visual quality
- dct domain
- computer vision
- high resolution
- wavelet coefficients
- low pass
- image blocks
- compressed images
- jpeg images
- image processing
- discrete wavelet transform
- image coding
- coded images
- wavelet domain
- frequency domain
- lapped transforms
- multiresolution
- similarity measure
- compression algorithm
- spatial domain
- high quality
- feature selection