# The Devil is in the Middle: Exploiting Mid-level Representations for Cross-Domain Instance Matching

arXiv: https://arxiv.org/abs/1711.08106 · HTML: https://mlsanity.com/paper/1711.08106/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Nov 2017.

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

- [Learning Disentangled Representation Implicitly via Transformer for Occluded Person Re-Identification](https://mlsanity.com/paper/2107.02380/) (2021)
- [HAT: Hierarchical Aggregation Transformers for Person Re-identification](https://mlsanity.com/paper/2107.05946/) (2021)
- [Pose-Guided Feature Learning with Knowledge Distillation for Occluded Person Re-Identification](https://mlsanity.com/paper/2108.00139/) (2021)
- [Connecting Language and Vision for Natural Language-Based Vehicle Retrieval](https://mlsanity.com/paper/2105.14897/) (2021)
- [Sampling Agnostic Feature Representation for Long-Term Person Re-identification](https://mlsanity.com/paper/2209.09574/) (2022)
- [UAV-ReID: A Benchmark on Unmanned Aerial Vehicle Re-identification in Video Imagery](https://mlsanity.com/paper/2104.06219/) (2021)
- [Training Deep Networks from Zero to Hero: avoiding pitfalls and going beyond](https://mlsanity.com/paper/2109.02752/) (2021)
- [Improving Person Re-Identification with Temporal Constraints](https://mlsanity.com/paper/2111.09056/) (2021)
- [Copy and Paste method based on Pose for Re-identification](https://mlsanity.com/paper/2107.10479/) (2021)
- [End-to-end Person Search Sequentially Trained on Aggregated Dataset](https://mlsanity.com/paper/2201.09604/) (2022)
- [Enhanced Vehicle Re-identification for ITS: A Feature Fusion approach using Deep Learning](https://mlsanity.com/paper/2208.06579/) (2022)
- [Is Synthetic Dataset Reliable for Benchmarking Generalizable Person Re-Identification?](https://mlsanity.com/paper/2209.05047/) (2022)
- [Learning Cross-modality Information Bottleneck Representation for Heterogeneous Person Re-Identification](https://mlsanity.com/paper/2308.15063/) (2023)

## Abstract (arXiv)

Many vision problems require matching images of object instances across different domains. These include fine-grained sketch-based image retrieval (FG-SBIR) and Person Re-identification (person ReID). Existing approaches attempt to learn a joint embedding space where images from different domains can be directly compared. In most cases, this space is defined by the output of the final layer of a deep neural network (DNN), which primarily contains features of a high semantic level. In this paper, we argue that both high and mid-level features are relevant for cross-domain instance matching (CDIM). Importantly, mid-level features already exist in earlier layers of the DNN. They just need to be extracted, represented, and fused properly with the final layer. Based on this simple but powerful idea, we propose a unified framework for CDIM. Instantiating our framework for FG-SBIR and ReID, we show that our simple models can easily beat the state-of-the-art models, which are often equipped with much more elaborate architectures.