• 4 min read
Enhancing Art Education at Kunstmuseum Bern Through a Digital Companion
Learn how museums like Kunstmuseum Bern are using computer vision to offer companion information in exhibits.
• 11 min read
Improving Vision Model Performance Using Roboflow & Tenyks
SUMMARY
A collaboration between Roboflow and Tenyks showing that model performance gains often come from fixing data, not tuning architecture. Using a traffic sign detection model as the test scenario, the workflow covers training a baseline, auditing the Roboflow dataset with Tenyks to surface label errors and class imbalances, correcting
• 8 min read
How to Use Kaggle for Computer Vision
In this guide, we show how to use Kaggle Notebooks for computer vision tasks.
• 7 min read
Automated Labeling for Images Organized in Folders
In this guide, we show how to use a nested folder-based structure to automatically label images with Autodistill.
• 7 min read
What is Data Augmentation? The Ultimate Guide.
In this guide, we talk about what data augmentation is, how augmented data can boost model performance, and how augmentations are used in computer vision.
• 6 min read
Ultimate Guide to Converting Bounding Boxes, Masks and Polygons
In this guide, we show how to convert bounding boxes (xyxy), masks, and polygons.
• 8 min read
How to Reduce Dataset Size Without Losing Accuracy
Learn how to reduce the size of your computer vision dataset without losing accuracy.
• 5 min read
Using Stable Diffusion and SAM to Modify Image Contents Zero Shot
SUMMARY
Combining Grounding DINO for zero-shot object detection, SAM for pixel-precise segmentation, and Stable Diffusion for inpainting produces a fully text-driven image editor that requires no manual masking or selection tools. The pipeline takes a text description of the object to modify, uses Grounding DINO to locate
• 10 min read
How to Use Computer Vision for Environment Monitoring
SUMMARY
Satellite and drone imagery analyzed by computer vision models can measure environmental change at a scale and frequency that manual scientific monitoring cannot match, turning pixel data into quantifiable ESG metrics like built-up area, deforestation extent, water coverage, and crop health. This guide walks through building a remote
• 5 min read
Detect and Segment Oil Spills Using Computer Vision
SUMMARY
Drone footage of an oil spill carries enough information to determine spill volume, thickness, and spatial extent, but only if the imagery is analyzed with a model trained to distinguish those properties. This post walks through building an instance segmentation model that labels oil spill regions by thickness class,
• 5 min read
How to Use LabelMe: A Complete Guide
In this guide, we discuss the features in LabelMe, how to install LabelMe, and how to start annotating images in the tool.
• 5 min read
Improve Accuracy: Polygon Annotations for Object Detection
SUMMARY
Polygon annotations consistently outperform bounding boxes for object detection when objects have irregular shapes, because polygons eliminate background noise that bounding boxes include. An experiment comparing both annotation types, combined with augmentations (rotation, saturation, cutout, bounding box shear) and pretrained weight initialization, showed meaningful mAP gains at each step.
• 5 min read
How to Use Generative AI to Augment Computer Vision Data
Dive deep into Kopikat.co, a tool for generative data augmentation created by OpenCV.ai to improve the quality of datasets.
• 6 min read
Train an Image Classification Model with No Labeling
In this guide, we show how to train an image classification model to identify damage on street signs without any labeling.
• 7 min read
Launch: Outsourced Data Labeling in Roboflow
Roboflow offers outsourced labeling via approved partners. Receive the custom annotations you need to train high-quality models.
• 2 min read
Launch: Adjust Image Brightness and Contrast in Roboflow Annotate
In this guide, we show how to use the brightness and contrast adjustment features available in Roboflow Annotate.
• 5 min read
Auto-Label Classification Datasets Using CLIP
SUMMARY
OpenAI's CLIP model can auto-label image classification datasets by computing cosine similarity between image embeddings and text embeddings for each class name, then assigning the best-matching label. This tutorial walks through the full pipeline in a Jupyter notebook using the Roboflow Art Classification Dataset, covering
• 8 min read
Not All mAPs are Equal and How to Test Model Robustness
Learn how to stress-test the robustness of computer vision models.
• 6 min read
Leveraging Embeddings and Clustering Techniques in Computer Vision
Explore the world of image embeddings in computer vision, as we dive into clustering, dataset assessment, and detecting image duplication. Discover dimensionality reduction techniques like t-SNE and UMAP. Use CLIP embeddings for analyzing image class distribution and identifying similar images.
• 5 min read
Zero-Shot Image Annotation with Grounding DINO and SAM - A Notebook Tutorial
In this comprehensive tutorial, discover how to speed up your image annotation process using Grounding DINO and Segment Anything Model. Learn how to convert object detection datasets into instance segmentation datasets, and use these models to automatically annotate your images.
• 4 min read
Launch: Label Data with Segment Anything in Roboflow
Learn how to label data with Segment Anything Model (SAM) in Roboflow.
• 5 min read
Synthetic Data Generation with NVIDIA and Roboflow
Learn how to build computer vision models that leverage synthetic data using NVIDIA Omniverse and Roboflow.
• 4 min read
How to Use Roboflow Models in CVAT
In this article, we show how to use public Roboflow models to speed up annotating in CVAT.
• 4 min read
Track and Count Objects Using YOLOv8
Counting moving objects is one of the most popular use cases in computer vision. It is used, among other things, in traffic analysis and as part of the automation of manufacturing processes. That is why understanding how to do it well is crucial for any CV engineer.
Stay Connected