CVAT Coco

## Overview

CVAT Coco is a dataset for object detection tasks - it contains Defect Distance Event annotations for 9,899 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Cvat

## Overview

Cvat is a dataset for computer vision tasks - it contains 1 annotations for 386 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

CVAT Upload

## Overview

CVAT Upload is a dataset for object detection tasks - it contains 123 annotations for 1,370 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Test Import From Cvat

## Overview

Test Import From Cvat is a dataset for object detection tasks - it contains Chess Test annotations for 269 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Overview

This dataset contains 8,992 images of Uno cards and 26,976 labeled examples on various textured backgrounds.

This dataset was collected, processed, and released by Roboflow user Adam Crawshaw, released with a modified MIT license: https://firstdonoharm.dev/

https://i.imgur.com/P8jIKjb.jpg" alt="Image example">

Use Cases

Adam used this dataset to create an auto-scoring Uno application:

Getting Started

Fork or download this dataset and follow our How to train state of the art object detector YOLOv4 for more.

Annotation Guide

See here for how to use the CVAT annotation tool.

About Roboflow

Roboflow makes managing, preprocessing, augmenting, and versioning datasets for computer vision seamless. :fa-spacer: Developers reduce 50% of their boilerplate code when using Roboflow's workflow, save training time, and increase model reproducibility. :fa-spacer:

Eye Detector 2 From CVAT

## Overview

Eye Detector 2 From CVAT is a dataset for object detection tasks - it contains Eyes annotations for 493 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [Public Domain license](https://creativecommons.org/licenses/Public Domain).

Industry 4.0 advancements offer promising solutions to the challenges faced by high-mix, low-volume (HMLV) foundries, particularly in quality assessment and process automation. This comprehensive dataset was developed to train an image segmentation neural network, aimed at automating the post-processing task of removing sprues and risers from cast parts. By enabling the analysis of diverse part geometries, the approach is designed to address the variability inherent in HMLV foundries, where standardization is difficult due to complex part configurations.

Data for this project consists of three types of images: camera images, synthetic images, and augmented images, all stored in JPEG format. Each sample contains 36 camera, synthetic, and augmented images. Camera images were captured using an Arduino Nicla Vision camera with a default resolution of 240 x 320 pixels. Images are labeled ‘Sample## Natural up/down ##’, with ## denoting the sample and image numbers. Synthetic images were created from raw 3D scan data and rendered in JPEG format using Blender at a default resolution of 1920 x 1080 pixels. They are labeled ‘Sample ## synthetic up/down ##’ Augmented images are synthetic images that have been modified by replacing the original part geometry with a CAD model. They are labeled ‘Sample ##A up/down ##’. The dataset includes both labeled and unlabeled images. The unlabeled set only includes JPEG images, while the labelled set includes JPEG images and their label in .txt format, as well as the .yaml file. A detailed description of the dataset creation is outlined below:

1) Real image creation: To create the real image dataset, each sample was placed on top of a turntable and a photo graph of the sample was taken. Then, the sample was rotated 20 degrees, and a subsequent photograph was taken. This process was repeated until a full rotation of the sample was complete, providing a total of 18 images. This process was then repeated with the object flipped upside down, providing 36 parts per sample, and 1080 images for the real dataset.

2) Synthetic image creation: The synthetic image dataset was created by using an Einscan Pro HD 3D scanner to collect 3D scans of the casted parts. The scans were imported into Blender and wrapped in an aluminum texture resembling the appearance real part. Then, the texture-wrapped part was place either in a blank scene with a black or gray background, or on top of a turntable resembling the real turntable in front of a white. Finally, the same image capture procedure performed on the real dataset was repeated in Blender to produce a total of 1080 synthetic images. All camera angles Fig. 4. Examples of augmented, real, and synthetic images from the dataset. and lighting were modelled to resemble the real images as closely as possible.

3) Augmented image creation: For the augmented image dataset, the same procedure for synthetic images was followed, with the exception of using Creo Parametic to replace the original part geometry with a CAD model of the part prior to importing into Blender. Similarly, 1080 Augmented parts were created.

Practitioners using this dataset have several options to tailor it to their needs. The dataset includes labeled and unlabeled images, with the labeled images organized into a single folder, enabling users to create custom train-validation-test splits. The unlabeled images can be categorized into different classes beyond the three provided in the labeled set (part, sprue, and riser). Additionally, the unlabeled set supports further 2D spatial augmentations, as class locations are specified in .txt files. The labeled images are named in the format ‘Sample XX-jpg.rf.RandomCharacterString,’ reflecting the dataset’s export from the data management platform Roboflow. Labeling can be performed in various platforms, such as Roboflow, CVAT, or Labelbox, providing flexibility in data management.

TREX2 CVAT Unreviewed

## Overview

TREX2 CVAT Unreviewed is a dataset for object detection tasks - it contains Rockets annotations for 525 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Food Anontation (CVAT Label)

## Overview

Food Anontation (CVAT Label) is a dataset for semantic segmentation tasks - it contains Food annotations for 882 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Training dataset

The SmartBay Observatory in Galway Bay is an important contribution by Ireland to the growing global network of real-time data capture systems deployed within the ocean – technology giving us new insights into the ocean which we have not had before.

The observatory was installed on the seafloor 1.5km off the coast of Spiddal, County Galway, Ireland . The observatory uses cameras, probes and sensors to permit continuous and remote live underwater monitoring. This observatory equipment allows ocean researchers unique real-time access to monitor ongoing changes in the marine environment. Data relating to the marine environment at the site is transferred in real-time from the SmartBay Observatory through a fibre optic telecommunications cable to the Marine Institute headquarters and onwards onto the internet. The data includes a live video stream, the depth of the observatory node, the sea temperature and salinity, and estimates of the chlorophyll and turbidity levels in the water which give an indication of the volume of phytoplankton and other particles, such as sediment, in the water.

The Smartbay Marine Species Object Detection training Dataset is an initial Bounding Box Annotated image dataset used in attempting to Train a YOLOv8 Object Detection Model to classify the Marine Fauna observed in the Smartbay Observatory Video footage using species names.

The imagery used in this training dataset consists of image frame captures from the Smartbay video Archive files, CC-BY imagery from the www.minka-sdg.org website and images taken by Eva Cullen in the "Galway Atlantaquaria" Aquarium in Galway, Ireland.

The imagery were annotated using CVAT, collated on Roboflow and exported in YOLOv8 training dataset format.

Butterfly Wing VIS CVAT

## Overview

Butterfly Wing VIS CVAT is a dataset for instance segmentation tasks - it contains Wing 9Y6m annotations for 582 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Overview

This dataset contains 74 images of aerial maritime photographs taken with via a Mavic Air 2 drone and 1,151 bounding boxes, consisting of docks, boats, lifts, jetskis, and cars. This is a multi class problem. This is an aerial object detection dataset. This is a maritime object detection dataset.

The drone was flown at 400 ft. No drones were harmed in the making of this dataset.

This dataset was collected and annotated by the Roboflow team, released with MIT license.

https://i.imgur.com/9ZYLQSO.jpg" alt="Image example">

Use Cases

• Identify number of boats on the water over a lake via quadcopter.
• Boat object detection dataset
• Aerial Object Detection proof of concept
• Identify if boat lifts have been taken out via a drone
• Identify cars with a UAV drone
• Find which lakes are inhabited and to which degree.
• Identify if visitors are visiting the lake house via quad copter.
• Proof of concept for UAV imagery project
• Proof of concept for maritime project
• Etc.

This dataset is a great starter dataset for building an aerial object detection model with your drone.

Getting Started

Fork or download this dataset and follow our How to train state of the art object detector YOLOv4 for more. Stay tuned for particular tutorials on how to teach your UAV drone how to see and comprable airplane imagery and airplane footage.

Annotation Guide

See here for how to use the CVAT annotation tool that was used to create this dataset.

About Roboflow

Roboflow makes managing, preprocessing, augmenting, and versioning datasets for computer vision seamless. :fa-spacer: Developers reduce 50% of their boilerplate code when using Roboflow's workflow, save training time, and increase model reproducibility. :fa-spacer:

CVAT_COCO_TO_OTHER

## Overview

CVAT_COCO_TO_OTHER is a dataset for computer vision tasks - it contains Caca annotations for 276 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

Summerschool_cvat

## Overview

Summerschool_cvat is a dataset for object detection tasks - it contains Flags annotations for 1,049 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Street_cvat

## Overview

Street_cvat is a dataset for object detection tasks - it contains Smoke annotations for 489 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Drone Image

## Overview

Drone Image is a dataset for object detection tasks - it contains Car Motorcycle Van Bus Lorry annotations for 9,029 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Basketball_v1_cvat

## Overview

Basketball_v1_cvat is a dataset for object detection tasks - it contains Basketball annotations for 538 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

This project is a product of the Hodin Lab at the University of Washington Friday Harbor Laboratories. It has the goal of developing a Pacific Northwest sea star instance segmentation and classification tool. Ultimately this tool will be used as a part of a sea star photo re-identification pipeline. Additionally, we hope this tool aids in the use of camera transect surveys of marine habitat.

The primary target species for this model is the Sunflower Seastar Pycnopodia helianthoides, the inclusion of other species is to make the model more robust to confusion species when deployed. For this reason, and our labs access to images of the Sunflower seastar it is over represented in the dataset.

If you have images of sea stars and wish to contribute to the project contact Willem @ willemlw@uw.edu

Our target number of annotated images is >10k with >100 annotated examples for each species.

We hope to have a future extension of this model which includes both star and prey annotations.

We are drawing images from a diverse set of sources including. iNaturalist Google search Collaborators Personal lab + field images boldsystems

We are using the annotation tool CVAT

Involved members are: Willem Lee Weertman Marilyn Duncan Ian Taylor Jason Hodin Brook Ashcraft

https://www.youtube.com/watch?v=4MA_6oZQz7s&ab_channel=tektronix475

Spotted caps, are the normal OK class (fully closed). Clean caps, are the bad or anomally target class (partially closed). One double prediction at 3:59. 100x100 classification accuracy, out of 200 samples. Inference over unseen test dataset. 150 epochs training. 700 samples training dataset, no data augmentation.

PREPROCESSING Auto-Orient: Applied Resize: Stretch to 416x416 Grayscale: Applied AUGMENTATIONS No augmentations were applied.

Anomaly detection with: Roboflow, tensorflow, google colab, Ultralytics, yolo v5, cvat,