VISEM-Tracking: human spermatozoa tracking video dataset

Provider
Vajira Thambawita et al., Simula Metropolitan Center for Digital Engineering and OsloMet
License
CC-BY-4.0 (source)
Size
121 files
Format
PNG, CSV (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/SimulaMet-HOST/VISEM-Tracking)
Access
Free and public. No account with the data provider is needed.
DOI
10.5281/zenodo.7293726
Category
Medicine
Source
https://arxiv.org/abs/2212.02842

About VISEM-Tracking: human spermatozoa tracking video dataset

VISEM-Tracking is a human spermatozoa tracking dataset published in Scientific Data (2023) by Simula and Oslo Metropolitan University. According to the paper it provides 20 video recordings of 30 seconds (29,196 frames) of wet sperm preparations under the microscope, with manually annotated bounding-box coordinates for every frame, a set of sperm characteristics analysed by domain experts, and additional unlabelled video clips for self-supervised or unsupervised learning. Boxes carry three classes: sperm, cluster, and small or pinhead. The authors report baseline detection with YOLOv5, and this repository includes their best YOLOv5 checkpoints. As measured 2026-11-14 the repository, uploaded by the authors' group, holds VISEM_tracking.zip (about 6.4 GB) plus a best_checkpoints folder; unzip to get the videos, extracted frames and YOLO-format label text files, readable with OpenCV and any YOLO toolkit. Used for sperm detection and multi-object tracking, motility and kinematics estimation, and andrology and fertility research. Released under CC BY 4.0 on Zenodo; cite Thambawita et al. 2023.

How to load VISEM-Tracking: human spermatozoa tracking video dataset in Python

import pandas as pd
df = pd.read_csv("hf://datasets/SimulaMet-HOST/VISEM-Tracking/best_checkpoints/yolov5l/results.csv")
df.head()

Data structure

File types in sampled files: PNG, CSV

Files

Example file: best_checkpoints/yolov5l/results.csv

Analyze this dataset without downloading it

On Scigantic, VISEM-Tracking: human spermatozoa tracking video dataset opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Medicine datasets