Object Detection On Train Track

## Overview

Object Detection On Train Track is a dataset for object detection tasks - it contains Person annotations for 50 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).

Pre-trained Object Detection Model found here: https://towardsdatascience.com/object-detection-with-10-lines-of-code-d6cb4d86f606

A RetinaNet model that I believe was trained using the COCO and Open Images datasets. Implied here: https://medium.com/deepquestai/train-object-detection-ai-with-6-lines-of-code-6d087063f6ff

Object detection is a vital part of any autonomous vision system and to obtain a high performing object detector data is needed. The object detection task aims to detect and classify different objects using camera input and getting bounding boxes containing the objects as output. This is usually done by utilizing deep neural networks.

When training an object detector a large amount of data is used, however it is not always practical to collect large amounts of data. This has led to multiple different techniques which decreases the amount of data needed. Examples of such techniques are transfer learning and domain adaptation. Working with construction equipment is a time consuming process and we wanted to examine if it was possible to use scale-model data to train a network and then used that network to detect real objects with no additional training.

This small dataset contains training and validation data of a scale dump truck in different environments while the test set contains images of a full size dump truck of similar model. The aim of the dataset is to train a network to classify wheels, cabs and tipping bodies of a scale-model dump truck and use that to classify the same classes on a full-scale dump truck.

The label structure of the dataset is the YOLO v3 structure, where the classes corresponds to a integer value, such that: Wheel: 0 Cab: 1 Tipping body: 2

Overview

The Hard Hat dataset is an object detection dataset of workers in workplace settings that require a hard hat. Annotations also include examples of just "person" and "head," for when an individual may be present without a hard hart.

The original dataset has a 75/25 train-test split.

Example Image: https://i.imgur.com/7spoIJT.png" alt="Example Image">

Use Cases

One could use this dataset to, for example, build a classifier of workers that are abiding safety code within a workplace versus those that may not be. It is also a good general dataset for practice.

Using this Dataset

Use the fork or Download this Dataset button to copy this dataset to your own Roboflow account and export it with new preprocessing settings (perhaps resized for your model's desired format or converted to grayscale), or additional augmentations to make your model generalize better. This particular dataset would be very well suited for Roboflow's new advanced Bounding Box Only Augmentations.

Dataset Versions:

Image Preprocessing | Image Augmentation | Modify Classes * v1 (resize-416x416-reflect): generated with the original 75/25 train-test split | No augmentations * v2 (raw_75-25_trainTestSplit): generated with the original 75/25 train-test split | These are the raw, original images * v3 (v3): generated with the original 75/25 train-test split | Modify Classes used to drop person class | Preprocessing and Augmentation applied * v5 (raw_HeadHelmetClasses): generated with a 70/20/10 train/valid/test split | Modify Classes used to drop person class * v8 (raw_HelmetClassOnly): generated with a 70/20/10 train/valid/test split | Modify Classes used to drop head and person classes * v9 (raw_PersonClassOnly): generated with a 70/20/10 train/valid/test split | Modify Classes used to drop head and helmet classes * v10 (raw_AllClasses): generated with a 70/20/10 train/valid/test split | These are the raw, original images * v11 (augmented3x-AllClasses-FastModel): generated with a 70/20/10 train/valid/test split | Preprocessing and Augmentation applied | 3x image generation | Trained with Roboflow's Fast Model * v12 (augmented3x-HeadHelmetClasses-FastModel): generated with a 70/20/10 train/valid/test split | Preprocessing and Augmentation applied, Modify Classes used to drop person class | 3x image generation | Trained with Roboflow's Fast Model * v13 (augmented3x-HeadHelmetClasses-AccurateModel): generated with a 70/20/10 train/valid/test split | Preprocessing and Augmentation applied, Modify Classes used to drop person class | 3x image generation | Trained with Roboflow's Accurate Model * v14 (raw_HeadClassOnly): generated with a 70/20/10 train/valid/test split | Modify Classes used to drop person class, and remap/relabel helmet class to head

Choosing Between Computer Vision Model Sizes | Roboflow Train

About Roboflow

Roboflow makes managing, preprocessing, augmenting, and versioning datasets for computer vision seamless.

Developers reduce 50% of their code when using Roboflow's workflow, automate annotation quality assurance, save training time, and increase model reproducibility.

A dataset of 80K+ construction site images sourced globally, featuring full EXIF data, including camera settings and photography details. Enriched with object and scene detection metadata, the dataset is ideal for AI model training in image recognition, classification, and segmentation.

Multi Instance Object Detection Dataset Sample

Duality.ai just released a 1000 image dataset used to train a YOLOv8 model for object detection -- and it's 100% free! Just create an EDU account here. This HuggingFace dataset is a 20 image and label sample, but you can get the rest at no cost by creating a FalconCloud account. Once you verify your email, the link will redirect you to the dataset page.

  Dataset Overview

This dataset consists of high-quality images of soup… See the full description on the dataset page: https://huggingface.co/datasets/duality-robotics/YOLOv8-Multi-Instance-Object-Detection-Dataset.

Vehicle Detection Dataset

This dataset is designed for vehicle detection tasks, featuring a comprehensive collection of images annotated for object detection. This dataset, originally sourced from Roboflow (https://universe.roboflow.com/object-detection-sn8ac/ai-traffic-system), was exported on May 29, 2025, at 4:59 PM GMT and is now publicly available on Kaggle under the CC BY 4.0 license.

Overview

• Purpose: The dataset supports the development of computer vision models for detecting various types of vehicles in traffic scenarios.
• Classes: The dataset includes annotations for 7 vehicle types:
  • Bicycle
  • Bus
  • Car
  • Motorbike
  • Rickshaw
  • Truck
  • Van
• Number of Images: The dataset contains 9,440 images, split into training, validation, and test sets:
  • Training: Images located in ../train/images
  • Validation: Images located in ../valid/images
  • Test: Images located in ../test/images
• Annotation Format: Images are annotated in YOLOv11 format, suitable for training state-of-the-art object detection models.
• Pre-processing: Each image has been resized to 640x640 pixels (stretched). No additional image augmentation techniques were applied.

Source and Creation

This dataset was created and exported via Roboflow, an end-to-end computer vision platform that facilitates collaboration, image collection, annotation, dataset creation, model training, and deployment. The dataset is part of the ai-traffic-system project (version 1) under the workspace object-detection-sn8ac. For more details, visit: https://universe.roboflow.com/object-detection-sn8ac/ai-traffic-system/dataset/1.

Usage

This dataset is ideal for researchers, data scientists, and developers working on vehicle detection and traffic monitoring systems. It can be used to: - Train and evaluate deep learning models for object detection, particularly using the YOLOv11 framework. - Develop AI-powered traffic management systems, autonomous driving applications, or urban mobility solutions. - Explore computer vision techniques for real-world traffic scenarios.

For advanced training notebooks compatible with this dataset, check out: https://github.com/roboflow/notebooks. To explore additional datasets and pre-trained models, visit: https://universe.roboflow.com.

License

The dataset is licensed under CC BY 4.0, allowing for flexible use, sharing, and adaptation, provided appropriate credit is given to the original source.

This dataset is a valuable resource for building robust vehicle detection models and advancing computer vision applications in traffic systems.

The Fish Detection AI project aims to improve the efficiency of fish monitoring around marine energy facilities to comply with regulatory requirements. Despite advancements in computer vision, there is limited focus on sonar images, identifying small fish with unlabeled data, and methods for underwater fish monitoring for marine energy.

A YOLO (You Only Look Once) computer vision model was developed using the Eyesea dataset (optical) and sonar images from Alaska Fish and Games to identify fish in underwater environments. Supervised methods were used within YOLO to detect fish based on training using labeled data of fish. These trained models were then applied to different unseen datasets, aiming to reduce the need for labeling datasets and training new models for various locations. Additionally, hyper-image analysis and various image preprocessing methods were explored to enhance fish detection.

In this research we achieved: 1. Enhanced YOLO Performance, as compared to a published article (Xu, Matzner 2018) using earlier yolo versions for fish object identification. Specifically, we achieved a best mean Average Precision (mAP) of 0.68 on the Eyesea optical dataset using YOLO v8 (medium-sized model), surpassing previous YOLO v3 benchmarks from that previous article publication. We further demonstrated up to 0.65 mAP on unseen sonar domains by leveraging a hyper-image approach (stacking consecutive frames), showing promising cross-domain adaptability.

This submission of data includes: - The actual best-performing trained YOLO model neural network weights, which can be applied to do object detection (PyTorch files, .pt). These are found in the Yolo_models_downloaded zip file - Documentation file to explain the upload and the goals of each of the experiments 1-5, as detailed in the word document (named "Yolo_Object_Detection_How_To_Document.docx") - Coding files, namely 5 sub-folders of python, shell, and yaml files that were used to run the experiments 1-5, as well as a separate folder for yolo models. Each of these is found in their own zip file, named after each experiment - Sample data structures (sample1 and sample2, each with their own zip file) to show how the raw data should be structured after running our provided code on the raw downloaded data - link to the article that we were replicating (Xu, Matzner 2018) - link to the Yolo documentation site from the original creators of that model (ultralytics) - link to the downloadable EyeSea data set from PNNL (instructions on how to download and format the data in the right way to be able to replicate these experiments is found in the How To word document)

Round 13 Train DatasetThis is the training data used to create and evaluate trojan detection software solutions. This data, generated at NIST, consists of object detection AIs trained both on synthetic image data build from Cityscapes and the DOTA_v2 dataset. A known percentage of these trained AI models have been poisoned with a known trigger which induces incorrect behavior. This data will be used to develop software solutions for detecting which trained AI models have been poisoned via embedded triggers. This dataset consists of 128 AI models using a small set of model architectures. Half (50%) of the models have been poisoned with an embedded trigger which causes misclassification of the input when the trigger is present.

Food Object Detection Train

## Overview

Food Object Detection Train is a dataset for object detection tasks - it contains Food annotations for 797 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).

The dataset is structured for person object detection tasks, containing separate directories for training, validation, and testing. Each split has an images folder with corresponding images and a labels folder with annotation files.

Train Set: Contains images and annotations for model training.

Validation Set: Includes images and labels for model evaluation during training.

Test Set: Provides unseen images and labels for final model performance assessment.

Each annotation file (TXT format) corresponds to an image and likely contains bounding box coordinates and class labels. This structure follows standard object detection dataset formats, ensuring easy integration with detection models like yolo,RT-DETR.

Dataset Structure

📂 dataset/ ├── 📁 train/ │ ├── 📂 images/ │ │ ├── 🖼 image1.jpg (Training image) │ │ ├── 🖼 image2.jpg (Training image) │ ├── 📂 labels/ │ │ ├── 📄 image1.txt (Annotation for image1.jpg) │ │ ├── 📄 image2.txt (Annotation for image2.jpg) │ ├── 📁 val/ │ ├── 📂 images/ │ │ ├── 🖼 image3.jpg (Validation image) │ │ ├── 🖼 image4.jpg (Validation image) │ ├── 📂 labels/ │ │ ├── 📄 image3.txt (Annotation for image3.jpg) │ │ ├── 📄 image4.txt (Annotation for image4.jpg) │ ├── 📁 test/ │ ├── 📂 images/ │ │ ├── 🖼 image5.jpg (Test image) │ │ ├── 🖼 image6.jpg (Test image) │ ├── 📂 labels/ │ │ ├── 📄 image5.txt (Annotation for image5.jpg) │ │ ├── 📄 image6.txt (Annotation for image6.jpg)

Round 10 Train DatasetThis is the training data used to create and evaluate trojan detection software solutions. This data, generated at NIST, consists of object detection AIs trained on the COCO dataset. A known percentage of these trained AI models have been poisoned with a known trigger which induces incorrect behavior. This data will be used to develop software solutions for detecting which trained AI models have been poisoned via embedded triggers. This dataset consists of 144 AI models using a small set of model architectures. Half (50%) of the models have been poisoned with an embedded trigger which causes misclassification of the input when the trigger is present.

A dataset of 1.9M+ traffic & road object images sourced globally, featuring full EXIF data, including camera settings and photography details. Enriched with object and scene detection metadata, the dataset is ideal for AI model training in image recognition, classification, and segmentation.

We purchased two packs of candy from a local store—a Nestlé “party pack” containing 60 “fun size” pieces of four candy types (Butterfinger, Crunch, Baby Ruth, and 100 Grand) and a Mars pack containing 125 “minis party size” pieces of five candy types (Snickers, Twix, Musketeers, Milky Way, and Milky Way Midnight). Given the nine candy types, there are a total of 126 unique combinations of four candy types. For each combination, we took eight photos, each containing four candy pieces, one for each type. The candy pieces were placed on a grey rug to improve the identification of candy pieces. Before taking each photo, we swapped candy pieces and rearranged their positions and sides. A Canon PowerShot SX540 digital camera was used to take photos. All images in the dataset were annotated using VGG Image Annotator (VIA). In an image, a rectangular bounding box was manually drawn around each candy piece with its type labeled.

A comprehensive dataset of 1M+ furniture images sourced globally, featuring full EXIF data, including camera settings and photography details. Enriched with object and scene detection metadata, this dataset is ideal for AI model training in image recognition, classification & segmentation

Object Detection Training

## Overview

Object Detection Training is a dataset for object detection tasks - it contains Helmet annotations for 1,000 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).

A comprehensive dataset of 10K+ package images sourced globally, featuring full EXIF data, including camera settings and photography details. Enriched with object and scene detection metadata, this dataset is ideal for AI model training in image recognition, classification, and segmentation.

This dataset features over 25,000,000 high-quality general-purpose images sourced from photographers worldwide. Designed to support a wide range of AI and machine learning applications, it offers a richly diverse and extensively annotated collection of everyday visual content.

Key Features: 1. Comprehensive Metadata: the dataset includes full EXIF data, detailing camera settings such as aperture, ISO, shutter speed, and focal length. Additionally, each image is pre-annotated with object and scene detection metadata, making it ideal for tasks like classification, detection, and segmentation. Popularity metrics, derived from engagement on our proprietary platform, are also included.

2.Unique Sourcing Capabilities: the images are collected through a proprietary gamified platform for photographers. Competitions spanning various themes ensure a steady influx of diverse, high-quality submissions. Custom datasets can be sourced on-demand within 72 hours, allowing for specific requirements—such as themes, subjects, or scenarios—to be met efficiently.

1. Global Diversity: photographs have been sourced from contributors in over 100 countries, covering a wide range of human experiences, cultures, environments, and activities. The dataset includes images of people, nature, objects, animals, urban and rural life, and more—captured across different times of day, seasons, and lighting conditions.

2. High-Quality Imagery: the dataset includes images with resolutions ranging from standard to high-definition to meet the needs of various projects. Both professional and amateur photography styles are represented, offering a balance of realism and creativity across visual domains.

3. Popularity Scores: each image is assigned a popularity score based on its performance in GuruShots competitions. This unique metric reflects how well the image resonates with a global audience, offering an additional layer of insight for AI models focused on aesthetics, engagement, or content curation.

4. AI-Ready Design: this dataset is optimized for AI applications, making it ideal for training models in general image recognition, multi-label classification, content filtering, and scene understanding. It integrates easily with leading machine learning frameworks and pipelines.

5. Licensing & Compliance: the dataset complies fully with data privacy regulations and offers transparent licensing for both commercial and academic use.

Use Cases: 1. Training AI models for general-purpose image classification and tagging. 2. Enhancing content moderation and visual search systems. 3. Building foundational datasets for large-scale vision-language models. 4. Supporting research in computer vision, multimodal AI, and generative modeling.

This dataset offers a comprehensive, diverse, and high-quality resource for training AI and ML models across a wide array of domains. Customizations are available to suit specific project needs. Contact us to learn more!

A diverse compilation of human facial images encompassing various races, age groups, and profiles, with the aim of creating an unbiased dataset that includes coordinates of facial regions suitable for training object detection models.

Buy me a coffee: https://bmc.link/baghbidi

Context

Training object-detection models on standard datasets can be quite computationally intensive. Here, I generate an object-detection dataset with MNIST to help learn and experiment more on the topic. This work is intended for those who want to try object detection with little computation resource.

A summary of an SSD model run on this data can be found here

Content

The data is generated from MNIST: several (<=4) digits are collected and put together to form an object-detection data sample. There are 16000 samples for training data and 2500 samples for testing data. Once loaded, the pickle files train.pkl or test.pkl contain a tuple of (images, boxes, labels).

Acknowledgements

Thank to my friend's suggestion and code

Inspiration

How do you improve detection of number "1"? 🤔