Multi-hop Deep Joint Source-Channel Coding with Deep Hash Distillation for Semantically Aligned Image Retrieval
This addresses secure image retrieval for multi-hop wireless communication, though it appears incremental as it combines existing DeepJSCC and hash distillation techniques.
The paper tackles image transmission over noisy multi-hop channels by training a DeepJSCC encoder-decoder with a deep hash distillation module to cluster images semantically, improving perceptual reconstruction quality. Results show significantly improved LPIPS scores compared to classical DeepJSCC, which suffers from noise accumulation.
We consider image transmission via deep joint source-channel coding (DeepJSCC) over multi-hop additive white Gaussian noise (AWGN) channels by training a DeepJSCC encoder-decoder pair with a pre-trained deep hash distillation (DHD) module to semantically cluster images, facilitating security-oriented applications through enhanced semantic consistency and improving the perceptual reconstruction quality. We train the DeepJSCC module to both reduce mean square error (MSE) and minimize cosine distance between DHD hashes of source and reconstructed images. Significantly improved perceptual quality as a result of semantic alignment is illustrated for different multi-hop settings, for which classical DeepJSCC may suffer from noise accumulation, measured by the learned perceptual image patch similarity (LPIPS) metric.