# Efficient Mirror Detection via Multi-level Heterogeneous Learning

arXiv: https://arxiv.org/abs/2211.15644 · HTML: https://mlsanity.com/paper/2211.15644/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 36 references that are in the map.
- First posted on arXiv 28 Nov 2022.

## Builds on (most cited first, 15 of 15)

- [Aggregated Residual Transformations for Deep Neural Networks](https://mlsanity.com/paper/1611.05431/) (2016)
- [Pyramid Scene Parsing Network](https://mlsanity.com/paper/1612.01105/) (2016)
- [Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers](https://mlsanity.com/paper/2012.15840/) (2020)
- [Cascaded Pyramid Network for Multi-Person Pose Estimation](https://mlsanity.com/paper/1711.07319/) (2017)
- [Coordinate Attention for Efficient Mobile Network Design](https://mlsanity.com/paper/2103.02907/) (2021)
- [EGNet:Edge Guidance Network for Salient Object Detection](https://mlsanity.com/paper/1908.08297/) (2019)
- [Cascaded Partial Decoder for Fast and Accurate Salient Object Detection](https://mlsanity.com/paper/1904.08739/) (2019)
- [Multi-scale Interactive Network for Salient Object Detection](https://mlsanity.com/paper/2007.09062/) (2020)
- [Visual Saliency Transformer](https://mlsanity.com/paper/2104.12099/) (2021)
- [Label Decoupling Framework for Salient Object Detection](https://mlsanity.com/paper/2008.11048/) (2020)
- [Global Context-Aware Progressive Aggregation Network for Salient Object Detection](https://mlsanity.com/paper/2003.00651/) (2020)
- [Learning Delicate Local Representations for Multi-Person Pose Estimation](https://mlsanity.com/paper/2003.04030/) (2020)
- [Direction-aware Spatial Context Features for Shadow Detection](https://mlsanity.com/paper/1712.04142/) (2017)
- [Blind Image Decomposition](https://mlsanity.com/paper/2108.11364/) (2021)
- [Symmetry-Aware Transformer-based Mirror Detection](https://mlsanity.com/paper/2207.06332/) (2022)

## Built on by (most cited first, 13 of 13)

- [SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation](https://mlsanity.com/paper/2408.08870/) (2024)
- [SegDINO: An Efficient Design for Medical and Natural Image Segmentation with DINO-V3](https://mlsanity.com/paper/2509.00833/) (2025)
- [Deep Learning in Concealed Dense Prediction](https://mlsanity.com/paper/2504.10979/) (2025)
- [LoD-Loc v2: Aerial Visual Localization over Low Level-of-Detail City Models using Explicit Silhouette Alignment](https://mlsanity.com/paper/2507.00659/) (2025)
- [SAM3-UNet: Simplified Adaptation of Segment Anything Model 3](https://mlsanity.com/paper/2512.01789/) (2025)
- [3DRef: 3D Dataset and Benchmark for Reflection Detection in RGB and Lidar Data](https://mlsanity.com/paper/2403.06538/) (2024)
- [A Racing Dataset and Baseline Model for Track Detection in Autonomous Racing](https://mlsanity.com/paper/2502.14068/) (2025)
- [Fusion of Short-term and Long-term Attention for Video Mirror Detection](https://mlsanity.com/paper/2407.07999/) (2024)
- [LSP-ST: Ladder Shape-Biased Side-Tuning for Robust Infrared Small Target Detection](https://mlsanity.com/paper/2504.14481/) (2025)
- [MirrorSAM2: Segment Mirror in Videos with Depth Perception](https://mlsanity.com/paper/2509.17220/) (2025)
- [MirrorMamba: Towards Scalable and Robust Mirror Detection in Videos](https://mlsanity.com/paper/2511.06716/) (2025)
- [Power of Boundary and Reflection: Semantic Transparent Object Segmentation using Pyramid Vision Transformer with Transparent Cues](https://mlsanity.com/paper/2512.07034/) (2025)
- [MVMD: A Multi-View Approach for Enhanced Mirror Detection](https://mlsanity.com/paper/2608.07559/) (2026)

## Abstract (arXiv)

We present HetNet (Multi-level Heterogeneous Network), a highly efficient mirror detection network. Current mirror detection methods focus more on performance than efficiency, limiting the real-time applications (such as drones). Their lack of efficiency is aroused by the common design of adopting homogeneous modules at different levels, which ignores the difference between different levels of features. In contrast, HetNet detects potential mirror regions initially through low-level understandings (textite.g., intensity contrasts) and then combines with high-level understandings (contextual discontinuity for instance) to finalize the predictions. To perform accurate yet efficient mirror detection, HetNet follows an effective architecture that obtains specific information at different stages to detect mirrors. We further propose a multi-orientation intensity-based contrasted module (MIC) and a reflection semantic logical module (RSL), equipped on HetNet, to predict potential mirror regions by low-level understandings and analyze semantic logic in scenarios by high-level understandings, respectively. Compared to the state-of-the-art method, HetNet runs 664% faster and draws an average performance gain of 8.9% on MAE, 3.1% on IoU, and 2.0% on F-measure on two mirror detection benchmarks.