DoubleQExt: Hardware and Memory Efficient CNN Through Two Levels of Quantization.
Jin-Chuan SeeHui-Fuang NgHung-Khoon TanJing-Jing ChangWai-Kong LeeSeong Oun HwangPublished in: IEEE Access (2021)
Keyphrases
- memory efficient
- real time
- low cost
- hardware and software
- multithreading
- iterative deepening
- external memory
- computer systems
- vlsi implementation
- integral image
- cellular neural networks
- image processing
- convolutional neural network
- hardware design
- levels of abstraction
- hardware implementation
- detection method
- multi dimensional
- multiresolution