CF-Net: A Cross-Feature Reconstruction Network for High-Accuracy 1-Bit Target Classification

TL;DR

CF-Net is a novel deep learning framework that reconstructs high-fidelity images from 1-bit radar data to enable high-accuracy target classification without oversampling.

Contribution

The paper introduces a dual-branch U-Net architecture with self-supervised pre-training for effective feature extraction from 1-bit radar data.

Findings

CF-Net achieves comparable or superior accuracy to 16-bit methods.

The framework effectively restores high-fidelity images from 1-bit data.

It enables high-accuracy classification without the need for oversampling.

Abstract

Target classification is a fundamental task in radar systems, and its performance critically depends on the quantization precision of the signal. While high-precision quantization (e.g. 16-bit) is well established, 1-bit quantization offers distinct advantages by enabling direct sampling at high frequencies and eliminating complex intermediate stages. However, its extreme quantization leads to significant information loss. Although higher sampling rates can compensate for this loss, such oversampling is impractical at the high frequencies targeted for direct sampling. To achieve high-accuracy classification directly from 1-bit radar data under the same sampling rate, this paper proposes a novel two-stage deep learning framework, CF-Net. First, we introduce a self-supervised pre-training strategy based on a dual-branch U-Net architecture. This network learns to restore high-fidelity 16-bit images from their 1-bit counterparts via a cross-feature reconstruction task, forcing the 1-bit encoder to learn robust features despite extreme quantization. Subsequently, this pre-trained encoder is repurposed and fine-tuned for the downstream multi-class target classification task. Experiments on two radar target datasets demonstrate that CF-Net can effectively extract discriminative features from 1-bit imagery, achieving comparable and even superior accuracy to some 16-bit methods without oversampling.