# textitA Contrario Paradigm for YOLO-based Infrared Small Target Detection

arXiv: https://arxiv.org/abs/2402.02288 · HTML: https://mlsanity.com/paper/2402.02288/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 20 references that are in the map.
- First posted on arXiv 3 Feb 2024.

## Builds on (most cited first, 11 of 11)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [You Only Look Once: Unified, Real-Time Object Detection](https://mlsanity.com/paper/1506.02640/) (2015)
- [Fast R-CNN](https://mlsanity.com/paper/1504.08083/) (2015)
- [YOLOv3: An Incremental Improvement](https://mlsanity.com/paper/1804.02767/) (2018)
- [YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors](https://mlsanity.com/paper/2207.02696/) (2022)
- [Asymmetric Contextual Modulation for Infrared Small Target Detection](https://mlsanity.com/paper/2009.14530/) (2020)
- [Dense Nested Attention Network for Infrared Small Target Detection](https://mlsanity.com/paper/2106.00487/) (2021)
- [You Only Learn One Representation: Unified Network for Multiple Tasks](https://mlsanity.com/paper/2105.04206/) (2021)
- [Efficient Evaluation of the Number of False Alarm Criterion](https://mlsanity.com/paper/1807.03594/) (2018)
- [Deep-NFA: a Deep textita contrario Framework for Small Object Detection](https://mlsanity.com/paper/2303.01363/) (2023)

## Built on by (most cited first, 2 of 2)

- [Robust infrared small target detection using self-supervised and a contrario paradigms](https://mlsanity.com/paper/2410.07437/) (2024)
- [ISTD-YOLO: A Multi-Scale Lightweight High-Performance Infrared Small Target Detection Algorithm](https://mlsanity.com/paper/2504.14289/) (2025)

## Abstract (arXiv)

Detecting small to tiny targets in infrared images is a challenging task in computer vision, especially when it comes to differentiating these targets from noisy or textured backgrounds. Traditional object detection methods such as YOLO struggle to detect tiny objects compared to segmentation neural networks, resulting in weaker performance when detecting small targets. To reduce the number of false alarms while maintaining a high detection rate, we introduce an textita contrario decision criterion into the training of a YOLO detector. The latter takes advantage of the textitunexpectedness of small targets to discriminate them from complex backgrounds. Adding this statistical criterion to a YOLOv7-tiny bridges the performance gap between state-of-the-art segmentation methods for infrared small target detection and object detection networks. It also significantly increases the robustness of YOLO towards few-shot settings.