Efficient Convolution Without Latency
نویسنده
چکیده
It is well known that a block FFT implementation of convolution is vastly more efficient than the direct form FIR filter. Unfortunately, block processing incurs significant input/output latency which is undesirable for real-time applications. A hybrid method is proposed for doing convolution by combining direct form and block FFT processing. The result is a zero latency convolver that performs significantly better than direct form methods.
منابع مشابه
Efficient Emotion Recognition from Speech Using Deep Learning on Spectrograms
We present a new implementation of emotion recognition from the para-lingual information in the speech, based on a deep neural network, applied directly to spectrograms. This new method achieves higher recognition accuracy compared to previously published results, while also limiting the latency. It processes the speech input in smaller segments – up to 3 seconds, and splits a longer input into...
متن کاملFace Recognition with Hybrid Efficient Convolution Algorithms on FPGAs
Deep Convolutional Neural Networks (CNN) have become a Swiss knife in solving critical artificial intelligence tasks. However, deploying deep CNN models for latency-critical tasks remains to be challenging because of the complex nature of CNNs. Recently, FPGA has become a favorable device to accelerate deep CNNs thanks to its high parallel processing capability and energy efficiency. In this wo...
متن کاملReducing Communication through Buffers on a Simd Architecture
Advances in wireless technology and the growing popularity of multimedia applications have brought about a need for energy efficient and cost effective portable supercomputers capable of delivering performance beyond the capabilities of current microprocessors and DSP chips. The SIMPil architecture currently being developed at Georgia Institute of Technology is a promising candidate for this ta...
متن کاملLatency-Optimized and Energy-Efficient MAC Protocol for Underwater Acoustic Sensor Networks: A Cross-Layer Approach
Considering the energy constraint for fixed sensor nodes and the unacceptable long propagation delay, especially for latency sensitive applications of underwater acoustic sensor networks, we propose a MAC protocol that is latency-optimized and energyefficient scheme and combines the physical layer and the MAC layer to shorten transmission delay. On physical layer, we apply convolution coding an...
متن کاملDesign and Implementation of Convolution Encoder and Viterbi Decoder
Data transmissions over wireless channels are affected by attenuation, distortion, interference and noise, which affect the receiver’s ability to receive correct information. Convolution encoding with Viterbi decoding is a powerful method for forward error correction. Convolution encoders and Viterbi decoders play an important role in digital communication especially, when channel is noisy and ...
متن کامل