# A Novel Lightweight Deep Learning Model for Boar Sperm Head Detection in Microscopic Images: YOLO11_SRP

## TL;DR

A new lightweight AI model called YOLO11_SRP improves the accuracy and efficiency of counting boar sperm heads in microscope images, helping with breeding and reproductive management.

## Contribution

A novel lightweight deep learning model, YOLO11_SRP, is introduced for efficient and accurate boar sperm head detection in microscopic images.

## Key findings

- YOLO11_SRP achieved an mAP@0.5 of 91.9%, a 13.9% improvement over YOLO11s.
- The model reduced parameters by 39% and computational cost by 14.1% compared to YOLO11s.

## Abstract

Accurately counting boar sperm heads is important for selecting high-quality breeding animals and improving reproductive efficiency on farms. Traditionally, workers must observe sperm under a microscope and count them by hand, which requires considerable time, effort, and experience. However, existing algorithms often have difficulty recognizing sperm cells when they overlap or exhibit high motility in high-magnification images, leading to unreliable results. In this study, we developed a new lightweight computer-based method that can automatically identify boar sperm heads in microscope images more efficiently and more accurately. This method uses an improved image-recognition model that focuses better on tiny objects and learns patterns from thousands of sperm images. When we tested the system, it achieved higher detection accuracy than a commonly used model while also requiring less computing power. The proposed method is intended as a foundational step for automated sperm analysis, providing reliable sperm head detection that can support downstream analysis in practical breeding applications.

Accurate and quantitative detection of boar sperm heads is essential for breeding selection and reproductive management. Manual microscopic counting is time-consuming, labor-intensive, and prone to subjective bias, while existing computer-based algorithms often struggle to recognize sperm cells accurately when they overlap or move rapidly in high-magnification microscopic images. This study proposes a lightweight boar sperm detection model, YOLO11_SRP, designed to improve small-object recognition in complex microscopic scenarios. The model integrates a lightweight StarNet backbone, a rectangular self-calibration module for enhanced spatial feature modeling, and an additional low-level detection layer optimized for tiny targets. We evaluated the model on a boar sperm microscopic image dataset and compared it with the standard YOLO11s framework. The results show that YOLO11_SRP achieves an mAP@0.5 of 91.9%, representing a 13.9% improvement over YOLO11s, while simultaneously reducing parameters by 39% and computational cost by 14.1%. These findings demonstrate that YOLO11_SRP provides efficient and accurate sperm detection, supporting the development of efficient and reliable automated sperm analysis pipelines, in which sperm head detection serves as a fundamental preprocessing step.