LLNet: A Deep Autoencoder approach to Natural Low-light Image Enhancement(利用深度自編碼器對低照度圖像進行增強)
一、自編碼器網絡結構
二、訓練過程
三、試驗結果
Comparison of methods of enhancing ‘Town’ when applied to (A) original already-bright, (B) darkened, (C) darkened and
noisy (σ = 18), and (D) darkened and noisy (σ = 25) images. Darkening is done with γ = 3. The numbers with units dB are PSNR,
the numbers without are SSIM. Best viewed on screen.
先貼幾張圖,後續在補過程!