• 4 min read
Track Football Players with Computer Vision
In this post, we use a YOLOv5 detection model and state-of-the-art tracker ByteTRACK to track football players.
• 7 min read
Preventing Accidents on Construction Sites with Computer Vision
In this article, we discuss how a community member built a system to prevent accidents on construction sites.
• 8 min read
Studying Links Between Litter and Socio-Economic Factors with Computer Vision
In this post, we talk about a report that studies links between litter and socio-economic factors with help from computer vision.
• 3 min read
Roboflow Changelog: November 2022
Each month, we share a recap of product and company updates. In this post, we detail Roboflow's updates from November 2022.
• 7 min read
What is a Segmentation Mask? An Introduction
In this guide, you will learn how to create segmentation masks with the results of a computer vision model hosted on Roboflow.
• 6 min read
Top 6 Environment Datasets for Computer Vision Projects
In this article, we highlight six open-source datasets you can use to solve various problems related to the environment with computer vision.
• 6 min read
📸 Roboflow 100: A Multi-Domain Object Detection Benchmark
Roboflow 100 (RF100) is a crowdsourced object detection benchmark. The dataset consists of 100 datasets, 7 imagery domains, 224,714 images, and 829 class labels with over 11,170 labeling hours.
• 14 min read
Turn Analog Dials into Digital Data with Computer Vision
In this article, we discuss how to turn information displayed on analog dials into digital data.
• 8 min read
What is a Confusion Matrix? A Beginner's Guide.
In this guide, we discuss what a confusion matrix is and how to use them to evaluate the performance of a computer vision model.
• 6 min read
5 Hobbyist Computer Vision Project Ideas
This article discusses five ideas of projects you could build to help you learn about computer vision.
• 15 min read
Monitoring My Caffeine Intake with Computer Vision
In this post, learn how to build a tool that monitors how many cups of tea or coffee you drink in a day.
• 4 min read
How to Use RPA to Supercharge Computer Vision Applications
SUMMARY
Robotic Process Automation tools like Microsoft Power Automate let you chain downstream actions onto computer vision inference results without writing a full backend. The post demonstrates the pattern by running inference on an image through a Roboflow model, sending the prediction data to a Power Automate webhook, and having
• 5 min read
Using Computer Vision to Save Sea Lions
SUMMARY
Steller sea lion numbers in Alaska's western Aleutian Islands have dropped 94% over 30 years, and NOAA researchers studying the cause generate thousands of remote camera images that must be reviewed manually. A computer vision tool called Steller Sight automates the triage step, classifying each image as
• 4 min read
Use Docker to Deploy Computer Vision Models
SUMMARY
Deploying a computer vision model involves far more than training it: managing Python packages, CUDA versions, system-level dependencies, and hardware differences all create friction before a single inference runs in production. Docker addresses this by packaging an application and all its dependencies into a portable container image, which
• 5 min read
Launch: Roboflow Command-line Interface
SUMMARY
The Roboflow CLI, installed via npm, lets you manage workspaces and projects, upload images and annotations, import datasets in common formats (COCO, Pascal VOC, and others), and run inference on object detection, classification, instance segmentation, or semantic segmentation models, all without opening a browser. Authentication stores API keys locally
• 5 min read
Top 6 Gaming Datasets for Computer Vision Projects
SUMMARY
Six gaming-oriented datasets from Roboflow Universe cover a range of real-world and digital game scenarios: a 350-image domino pip dataset for score tracking, a 6.3K-image Counter-Strike player dataset with CT and T team labels, draughts and chess piece detection sets, a billiards ball
• 5 min read
Top 7 Sports Datasets for Computer Vision Projects
SUMMARY
Seven sports datasets on Roboflow Universe span classification and object detection across multiple disciplines: cricket, football, and baseball image classification; a 5K-image football helmet detection set for impact tracking; shuttlecock and golf ball detection for racket and club sports; a boxing punch detector; a cyclist database; and a
• 8 min read
Identifying Objects in Multimedia Databases with Computer Vision
In this post, we discuss research from Prof. Luigi Laura and Mirko Cangiano showing how to identify objects in multimedia databases.
• 11 min read
Synthetic Data Generation with Stable Diffusion: A Guide
In this tutorial, we walk through how to generate images with Stable Diffusion for use in a computer vision model.
• 7 min read
Detecting and Reading QR Codes Using Computer Vision
In this article, we're going to walk through how to detect and read QR codes using Roboflow and Python.
• 4 min read
Roboflow Changelog: October 2022
Each month, we share a recap of product and company updates. In this post, we detail Roboflow's updates from October 2022.
• 7 min read
Overfitting in Machine Learning and Computer Vision
SUMMARY
Overfitting happens when a model memorizes training data instead of learning general patterns, a problem visible when training accuracy is high but test accuracy drops sharply. Detecting it requires monitoring the gap between training and validation performance across epochs, and preventing it draws on techniques including data augmentation, regularization,
• 5 min read
Top 6 Agriculture Datasets for Computer Vision
Explore six agriculture datasets you can use in computer vision projects.
• 3 min read
Using AI to Count Calories from Photos with SnapCalorie
SUMMARY
SnapCalorie built a mobile app that photographs a meal and returns estimated calories, fat, carbs, and protein, a task that demands fast iteration on computer vision models as the food dataset grows. By adopting Roboflow Annotate for dataset management, labeling, annotation review, and model-assisted labeling, the team increased
Stay Connected