<p>Low-light image enhancement remains a challenging task due to severe luminance degradation, color distortion, and detail loss. This paper proposes FCE-Net, a novel YUV color space decomposition network that jointly leverages Frequency, Color, and Edge enhancement modules to address these challenges. The input image is first transformed into the YUV space to decouple luminance and chrominance information. For the chrominance (U, V) channels, an attention-guided dual-branch denoising module effectively suppresses color noise and preserves color consistency. For the luminance (Y) channel, a high-low frequency separation module extracts structural and textural details, which are further refined by an edge enhancement block and multi-head self-attention to strengthen feature representation. A multi-stage squeeze and excitation fusion module adaptively integrates cross-scale features, ensuring robust restoration under diverse lighting conditions. Extensive experiments on ten public datasets demonstrate that FCE-Net achieves excellent performance in lightweight models that use only 0.68M parameters, delivering superior perceptual quality and computational efficiency. The source code is publicly available at <a href="https://github.com/gzy7/FCE-Net">https://github.com/gzy7/FCE-Net</a>.</p>