LSENet: A Lightweight Spectral Enhancement Network for High-Quality Speech Processing on Resource-Constrained Platforms
Hyeong Il Koh, Sungdae Na, Myoung Nam Kim · IEEE Access · 2025
Although recent deep-learning-based speech enhancement (SE) methods significantly outperform traditional approaches, their computational demands often scale proportionally with their performance. This makes them usually impractical for deployment on resource-constrained computing platforms (edge devices) like smartphones, IoT devices, and digital hearing aids, which process data locally. In this paper, we propose a novel lightweight spectral enhancement network (LSENet) that is designed to estimate high-quality speech with minimal computational overhead. The network consists of an encoder-decoder architecture enhanced by a group-dilated convolutional module, which efficiently leverages time-frequency domain information while significantly reducing resource consumption through dilated convolutional groups and spectral-wise attention modules. In addition, to capture the long-range contextual dependencies of the extracted features, an improved dual-path recurrent neural network is introduced between the encoder-decoder structure. Experimental results show that the proposed model achieves competitive performance with state-of-the-art baseline models on the Voicebank + Demand and DNS-Challenge datasets while requiring only 39.4 thousand model parameters and 237 million multiply-accumulate operations.