# Efficient INT8 Single-Image Super-Resolution via Deployment-Aware Quantization and Teacher-Guided Training

arXiv: https://arxiv.org/abs/2604.20291 · HTML: https://mlsanity.com/paper/2604.20291/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 33 references that are in the map.
- First posted on arXiv 22 Apr 2026.

## Builds on (most cited first, 27 of 27)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [SwinIR: Image Restoration Using Swin Transformer](https://mlsanity.com/paper/2108.10257/) (2021)
- [Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference](https://mlsanity.com/paper/1712.05877/) (2017)
- [Enhanced Deep Residual Networks for Single Image Super-Resolution](https://mlsanity.com/paper/1707.02921/) (2017)
- [Image Super-Resolution Using Very Deep Residual Channel Attention Networks](https://mlsanity.com/paper/1807.02758/) (2018)
- [Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network](https://mlsanity.com/paper/1609.05158/) (2016)
- [RepVGG: Making VGG-style ConvNets Great Again](https://mlsanity.com/paper/2101.03697/) (2021)
- [Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution](https://mlsanity.com/paper/1704.03915/) (2017)
- [Activating More Pixels in Image Super-Resolution Transformer](https://mlsanity.com/paper/2205.04437/) (2022)
- [Accelerating the Super-Resolution Convolutional Neural Network](https://mlsanity.com/paper/1608.00367/) (2016)
- [MambaIR: A Simple Baseline for Image Restoration with State-Space Model](https://mlsanity.com/paper/2402.15648/) (2024)
- [Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network](https://mlsanity.com/paper/1803.08664/) (2018)
- [Lightweight Image Super-Resolution with Information Multi-distillation Network](https://mlsanity.com/paper/1909.11856/) (2019)
- [Residual Feature Distillation Network for Lightweight Image Super-Resolution](https://mlsanity.com/paper/2009.11551/) (2020)
- [Fast and Accurate Single Image Super-Resolution via Information Distillation Network](https://mlsanity.com/paper/1803.09454/) (2018)
- [MobileOne: An Improved One millisecond Mobile Backbone](https://mlsanity.com/paper/2206.04040/) (2022)
- [MambaIRv2: Attentive State Space Restoration](https://mlsanity.com/paper/2411.15269/) (2024)
- [Confidence-Aware Multi-Teacher Knowledge Distillation](https://mlsanity.com/paper/2201.00007/) (2021)
- [Anchor-based Plain Net for Mobile Image Super-Resolution](https://mlsanity.com/paper/2105.09750/) (2021)
- [DVMSR: Distillated Vision Mamba for Efficient Super-Resolution](https://mlsanity.com/paper/2405.03008/) (2024)
- [Efficient and Accurate Quantized Image Super-Resolution on Mobile NPUs, Mobile AI & AIM 2022 challenge: Report](https://mlsanity.com/paper/2211.05910/) (2022)
- [Improving Generalization in Visual Reasoning via Self-Ensemble](https://mlsanity.com/paper/2410.20883/) (2024)
- [FreqNet: A Frequency-domain Image Super-Resolution Network with Dicrete Cosine Transform](https://mlsanity.com/paper/2111.10800/) (2021)
- [STER-VLM: Spatio-Temporal With Enhanced Reference Vision-Language Models](https://mlsanity.com/paper/2508.13470/) (2025)
- [Hybrid, Unified and Iterative: A Novel Framework for Text-based Person Anomaly Retrieval](https://mlsanity.com/paper/2511.22470/) (2025)
- [Post-Training BatchNorm Recalibration](https://mlsanity.com/paper/2010.05625/) (2020)
- [Cycle Training with Semi-Supervised Domain Adaptation: Bridging Accuracy and Efficiency for Real-Time Mobile Scene Detection](https://mlsanity.com/paper/2504.09297/) (2025)

## Abstract (arXiv)

Efficient single-image super-resolution (SISR) requires balancing reconstruction fidelity, model compactness, and robustness under low-bit deployment, which is especially challenging for x3 SR. We present a deployment-oriented quantized SISR framework based on an extract-refine-upsample design. The student performs most computation in the low-resolution space and uses a lightweight re-parameterizable backbone with PixelShuffle reconstruction, yielding a compact inference graph. To improve quality without significantly increasing complexity, we adopt a three-stage training pipeline: Stage 1 learns a basic reconstruction mapping with spatial supervision; Stage 2 refines fidelity using Charbonnier loss, DCT-domain supervision, and confidence-weighted output-level distillation from a Mamba-based teacher; and Stage 3 applies quantization-aware training directly on the fused deploy graph. We further use weight clipping and BatchNorm recalibration to improve quantization stability. On the MAI 2026 Quantized 4K Image Super-Resolution Challenge test set, our final AIO MAI submission achieves 29.79 dB PSNR and 0.8634 SSIM, obtaining a final score of 1.8 under the target mobile INT8 deployment setting. Ablation on Stage 3 optimization shows that teacher-guided supervision improves the dynamic INT8 TFLite reconstruction from 29.91 dB/0.853 to 30.0003 dB/0.856, while the fixed-shape deployable INT8 TFLite artifact attains 30.006 dB/0.857.