TY  - CONF
AU  - Guillem Martinez
AU  - Maya Aghaei
AU  - Martin Dijkstra
AU  - Bhalaji Nagarajan
AU  - Femke Jaarsma
AU  - Jaap van de Loosdrecht
AU  - Petia Radeva
AU  - Klaas Dijkstra
A2  - ICASSP
PY  - 2022//
TI  - Hyper-Spectral Imaging for Overlapping Plastic Flakes Segmentation
BT  - 47th International Conference on Acoustics, Speech, and Signal Processing
KW  - Hyper-spectral imaging
KW  - plastic sorting
KW  - multi-label segmentation
KW  - bitfield encoding
N2  - In this paper, we propose a deformable convolution-based generative adversarial network (DCNGAN) for perceptual quality enhancement of compressed videos. DCNGAN is also adaptive to the quantization parameters (QPs). Compared with optical flows, deformable convolutions are more effective and efficient to align frames. Deformable convolutions can operate on multiple frames, thus leveraging more temporal information, which is beneficial for enhancing the perceptual quality of compressed videos. Instead of aligning frames in a pairwise manner, the deformable convolution can process multiple frames simultaneously, which leads to lower computational complexity. Experimental results demonstrate that the proposed DCNGAN outperforms other state-of-the-art compressed video quality enhancement algorithms.
UR  - https://ieeexplore.ieee.org/document/9897749
L1  - http://refbase.cvc.uab.es/files/MAD2022.pdf
UR  - http://dx.doi.org/10.1109/ICIP46576.2022.9897749
N1  - MILAB; no proj
ID  - Guillem Martinez2022
ER  -