Drowning People

## Overview

Drowning People is a dataset for object detection tasks - it contains Drowning People annotations for 93 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).

titanic5 Dataset Created by David Beltran del Rio March 2016.

Notes This is the final (for now) version of my update to the Titanic data. I think it’s finally ready for publishing if you’d like. What I did was to strip all the passenger and crew data from the Encyclopedia Titanica (ET) web pages (excluding channel crossing passengers), create a unique ID for each passenger and crew member (Name_ID), then (painstakingly and hopefully 100% correctly) match to your earlier titanic3 dataset, in order to compare the two and to get your sibsp and parch variables. Since the ET is updated occasionally the work put into the ID and matching can be reused and refined later. I did eventually hear back from the ET people, they are willing to make the underlying database available in the future, I have not yet taken them up on it.

The two datasets line up nicely, most of the differences in the newer titanic5 dataset are in the age variable, as I had mentioned before - the new set has less missing ages - 51 missing (vs 263) out of 1309.

I am in the process of refining my analysis of the data as well, based on your comments below and your Regression Modeling Strategies example.

titanic3_wID data can be matched to titanic5 using the Name_ID variable. Tab titanic5 Metadata has the variable descriptions and allowable values for Class and Class/Dept.

A note about the ages - instead of using the add 0.5 trick to indicate estimated birth day / date I have a flag that indicates how the “final” age (Age_F) was arrived at. It’s the Age_F_Code variable - the allowable values are in the Titanic5_metadata tab in the attached excel. The reason for this is that I already had some fractional ages for infants where I had age in months instead of years and I wanted to avoid confusion for 6 month old infants, although I don’t think there are any in the data! Also, I was thinking to make fractional ages or age in days for all passengers for whom I have DoB, but I have not yet done so.

Here’s what the tabs are:

Titanic5_all - all (mostly cleaned) Titanic passenger and crew records Titanic5_work - working dataset, crew removed, unnecessary variables removed - this is the one I import into SAS / R to work on Titanic5_metadata - Variable descriptions and allowable values titanic3_wID - Original Titanic3 dataset with Name_ID added for merging to Titanic5 I have a csv, R dataset, and SAS dataset, but the variable names are an older version, so I won’t send those along for now to avoid confusion.

If it helps send my contact info along to your student in case any questions arise. Gmail address probably best, on weekends for sure: davebdr@gmail.com

The tabs in titanic5.xls are

Titanic5_all Titanic5_passenger (the one to be used for analysis) Titanic5_metadata (used during analysis file creation) Titanic3_wID

Here are a few use cases for this project:

1. Human Presence Detection: This computer vision model can be incorporated into security systems and smart home devices to identify the presence of humans in an area, allowing for customized responses, room automation, and improved safety.

2. Crowd Size Estimation: The "human dataset v1" can be used by event organizers or city planners to estimate the size of gatherings or crowds at public events, helping them better allocate resources and manage these events more efficiently.

3. Surveillance and Security Enhancement: The model can be integrated into video surveillance systems to more accurately identify humans, helping to filter out false alarms caused by animals and other non-human entities.

4. Collaborative Robotics: Robots equipped with this computer vision model can more easily identify and differentiate humans from their surroundings, allowing them to more effectively collaborate with people in shared spaces while ensuring human safety.

5. Smart Advertising: The "human dataset v1" can be utilized by digital signage and advertising systems to detect and count the number of human viewers, enabling targeted advertising and measuring the effectiveness of marketing campaigns.

Description

The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) contains 7356 files (total size: 24.8 GB). The dataset contains 24 professional actors (12 female, 12 male), vocalizing two lexically-matched statements in a neutral North American accent. Speech includes calm, happy, sad, angry, fearful, surprise, and disgust expressions, and song contains calm, happy, sad, angry, and fearful emotions. Each expression is produced at two levels of emotional intensity (normal, strong), with an additional neutral expression. All conditions are available in three modality formats: Audio-only (16bit, 48kHz .wav), Audio-Video (720p H.264, AAC 48kHz, .mp4), and Video-only (no sound). Note, there are no song files for Actor_18.

The RAVDESS was developed by Dr Steven R. Livingstone, who now leads the Affective Data Science Lab, and Dr Frank A. Russo who leads the SMART Lab.

Citing the RAVDESS

The RAVDESS is released under a Creative Commons Attribution license, so please cite the RAVDESS if it is used in your work in any form. Published academic papers should use the academic paper citation for our PLoS1 paper. Personal works, such as machine learning projects/blog posts, should provide a URL to this Zenodo page, though a reference to our PLoS1 paper would also be appreciated.

Academic paper citation

Livingstone SR, Russo FA (2018) The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS): A dynamic, multimodal set of facial and vocal expressions in North American English. PLoS ONE 13(5): e0196391. https://doi.org/10.1371/journal.pone.0196391.

Personal use citation

Include a link to this Zenodo page - https://zenodo.org/record/1188976

Commercial Licenses

Commercial licenses for the RAVDESS can be purchased. For more information, please visit our license page of fees, or contact us at ravdess@gmail.com.

Contact Information

If you would like further information about the RAVDESS, to purchase a commercial license, or if you experience any issues downloading files, please contact us at ravdess@gmail.com.

Example Videos

Watch a sample of the RAVDESS speech and song videos.

Emotion Classification Users

If you're interested in using machine learning to classify emotional expressions with the RAVDESS, please see our new RAVDESS Facial Landmark Tracking data set [Zenodo project page].

Construction and Validation

Full details on the construction and perceptual validation of the RAVDESS are described in our PLoS ONE paper - https://doi.org/10.1371/journal.pone.0196391.

The RAVDESS contains 7356 files. Each file was rated 10 times on emotional validity, intensity, and genuineness. Ratings were provided by 247 individuals who were characteristic of untrained adult research participants from North America. A further set of 72 participants provided test-retest data. High levels of emotional validity, interrater reliability, and test-retest intrarater reliability were reported. Validation data is open-access, and can be downloaded along with our paper from PLoS ONE.

Contents

Audio-only files

Audio-only files of all actors (01-24) are available as two separate zip files (~200 MB each):

• Speech file (Audio_Speech_Actors_01-24.zip, 215 MB) contains 1440 files: 60 trials per actor x 24 actors = 1440.
• Song file (Audio_Song_Actors_01-24.zip, 198 MB) contains 1012 files: 44 trials per actor x 23 actors = 1012.

Audio-Visual and Video-only files

Video files are provided as separate zip downloads for each actor (01-24, ~500 MB each), and are split into separate speech and song downloads:

• Speech files (Video_Speech_Actor_01.zip to Video_Speech_Actor_24.zip) collectively contains 2880 files: 60 trials per actor x 2 modalities (AV, VO) x 24 actors = 2880.
• Song files (Video_Song_Actor_01.zip to Video_Song_Actor_24.zip) collectively contains 2024 files: 44 trials per actor x 2 modalities (AV, VO) x 23 actors = 2024.

File Summary

In total, the RAVDESS collection includes 7356 files (2880+2024+1440+1012 files).

File naming convention

Each of the 7356 RAVDESS files has a unique filename. The filename consists of a 7-part numerical identifier (e.g., 02-01-06-01-02-01-12.mp4). These identifiers define the stimulus characteristics:

Filename identifiers

• Modality (01 = full-AV, 02 = video-only, 03 = audio-only).
• Vocal channel (01 = speech, 02 = song).
• Emotion (01 = neutral, 02 = calm, 03 = happy, 04 = sad, 05 = angry, 06 = fearful, 07 = disgust, 08 = surprised).
• Emotional intensity (01 = normal, 02 = strong). NOTE: There is no strong intensity for the 'neutral' emotion.
• Statement (01 = "Kids are talking by the door", 02 = "Dogs are sitting by the door").
• Repetition (01 = 1st repetition, 02 = 2nd repetition).
• Actor (01 to 24. Odd numbered actors are male, even numbered actors are female).

Filename example: 02-01-06-01-02-01-12.mp4

1. Video-only (02)
2. Speech (01)
3. Fearful (06)
4. Normal intensity (01)
5. Statement "dogs" (02)
6. 1st Repetition (01)
7. 12th Actor (12)
8. Female, as the actor ID number is even.

License information

The RAVDESS is released under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, CC BY-NC-SA 4.0

Commercial licenses for the RAVDESS can also be purchased. For more information, please visit our license fee page, or contact us at ravdess@gmail.com.

Related Data sets

• RAVDESS Facial Landmark Tracking data set [Zenodo project page].

CoLA: The Corpus of Linguistic Acceptability

Introduction

The Corpus of Linguistic Acceptability (CoLA) in its full form consists of 10657 sentences from 23 linguistics publications, expertly annotated for acceptability (grammaticality) by their original authors. The public version provided here contains 9594 sentences belonging to training and development sets, and excludes 1063 sentences belonging to a held out test set. Contact alexwarstadt [at] gmail [dot] com with any questions or issues. Read the paper or check out the source code for baselines.

Acknowledgement

from https://nyu-mll.github.io/CoLA/

A dataset of English Twitter messages with six basic emotions: anger, fear, joy, love, sadness, and surprise. For more detailed information please refer to the paper below.

The authors constructed a set of hashtags to collect a separate dataset of English tweets from the Twitter API belonging to eight basic emotions, including anger, anticipation, disgust, fear, joy, sadness, surprise, and trust. The data has already been preprocessed based on the approach described in their paper.

• Homepage: https://github.com/dair-ai/emotion_dataset
• Paper: CARER: Contextualized Affect Representations for Emotion Recognition
• Point of Contact: ellfae@gmail.com
• Size of downloaded dataset files: 3.95 MB
• Size of the generated dataset: 4.16 MB
• Total amount of disk used: 8.11 MB

An example of 'train' looks as follows. { "label": 0, "text": "im feeling quite sad and sorry for myself but ill snap out of it soon" }

Starter Notebook

Exploratory Data Analysis of the emotion dataset

Acknowledgements

@inproceedings{saravia-etal-2018-carer,
  title = "{CARER}: Contextualized Affect Representations for Emotion Recognition",
  author = "Saravia, Elvis and
   Liu, Hsien-Chi Toby and
   Huang, Yen-Hao and
   Wu, Junlin and
   Chen, Yi-Shin",
  booktitle = "Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing",
  month = oct # "-" # nov,
  year = "2018",
  address = "Brussels, Belgium",
  publisher = "Association for Computational Linguistics",
  url = "https://www.aclweb.org/anthology/D18-1404",
  doi = "10.18653/v1/D18-1404",
  pages = "3687--3697",
  abstract = "Emotions are expressed in nuanced ways, which varies by collective or individual experiences, knowledge, and beliefs. Therefore, to understand emotion, as conveyed through text, a robust mechanism capable of capturing and modeling different linguistic nuances and phenomena is needed. We propose a semi-supervised, graph-based algorithm to produce rich structural descriptors which serve as the building blocks for constructing contextualized affect representations from text. The pattern-based representations are further enriched with word embeddings and evaluated through several emotion recognition tasks. Our experimental results demonstrate that the proposed method outperforms state-of-the-art techniques on emotion recognition tasks.",
}

Handicapped people are a group of Internet and computer users. Some assistive technologies are designed for these people depending on their voice to apply voice commands to use technology. Gmail is one of the freely available and popular e-mail services and it has currently become most of people using Gmail as communication way with each other. The main aim of this paper is to introduce a computer application called HPG that can control and navigation the Gmail by voice commands. H_P_G helps handicapped people to send and receive their own emails, easily login and logout to their Gmail accounts, attach files if desired, and send and receive emails. The advantages of the developed system are that it is very low-cost, easy to use, and helps handicapped people to be in contact with their friends and family using their Gmail accounts. The developed application should also be useful to other researchers who wish to develop computer based applications for the handicapped people.

Here are a few use cases for this project:

1. Education and Learning Applications: This model could be used in interactive learning apps or software to create a number-learning game for children. By identifying numbers displayed in real-time, it transfers abstract learning to a more engaging, interactive experience.

2. Document Analysis and Data Extraction: Companies dealing with large volumes of printed numbers or documents with numerous numerical figures could use this model to automatically extract and convert these numbers into digital format.

3. Real Estate Inventory Management: The model can be utilized for recognizing house numbers from images or street view data. This can help maintain a structured inventory of real estate properties and streamline real estate operations.

4. Retail: It can be employed in retail store management, especially in inventory control, by recognizing product numbers on their labels, thus automating the inventory update process.

5. Assistive Technology: Develop a system that identifies numerical data in the environment for visually impaired people, helping them navigate daily tasks more independently.

Narratives: fMRI data for evaluating models of naturalistic language comprehension

The "Narratives" collection aggregates auditory story-listening fMRI datasets acquired over the course of roughly seven years (2011–2018). Stimuli comprised 28 naturalistic spoken stories ranging from ~3 to ~56 minutes for a total of ~5 hours of unique audio stimuli. The collection includes 345 unique subjects participating in over 750 functional scans with accompanying anatomical data. This re-release of the dataset follows on ds002245 v.1.0.3 and fixes some issues with cropped and redundant T1w anatomical images.

Please contact Sam Nastase if you observe any irregularities in the dataset.

Samuel A. Nastase sam.nastase@gmail.com

Validationset Conversion

## Overview

Validationset Conversion is a dataset for object detection tasks - it contains People annotations for 2,487 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).

Our aim is to identify Hand Gesture from the given image and display the result in text format or audio which will be useful for Hearing impaired people. T0 train the CNN model, we have prepared our own dataset. The following are the dataset details

Image Resolution : 12 mega pixel Image Size : 1920 * 1080

The dataset has 9 hand gestures. The following are the hand gestures:

Class ID : Class Name "1": "Have", "2": "Nice", "3": "Day", "4": "Early", "5": "Morning", "6": "Wakeup", "7": "Love", "8": "Funny", "9": "You"

Train dataset has 232 images and Validation Dataset has 55 images.

All the images are annotated using VGG Annotator tool .

Annotation Details:

Hand Gesture is annotated with polygon coordinates. Annotated only the hand region (Palm and Fingers). Annotation information are stored in JSON file (via_region_annotation.json)

Each Hand Gesture has 20 images in Train Dataset and 5 images in validation Dataset.

In our Project, we have used MASK RCNN to detect the Hand Gesture . It gives 3 results such as Class Name, Bounding Box Regressor and Segmentation.

Accuracy Score : Intersection Over Union (IoU) - 0.875 and mAP (Mean Arithmetic Precision) - 0.95

If you have any queries. Please reach out to us via email (HSL.Queries@gmail.com)

About the dataset: Hurricane Michael was the third-most intense Atlantic hurricane to make landfall in the United States in terms of pressure. This dataset was collected from Twitter during Hurricane Michael. The dataset was processed and analyzed using the AIDR (http://aidr.qcri.org) platform.

Dataset Description: This is a Twitter dataset collected during Hurricane Michael 2018. The data was collected, processed, and analyzed by the AIDR (http://aidr.qcri.org) platform using state-of-the-art machine learning techniques. The data includes the number of injured and dead people, infrastructure damage reports, missing or found people, urgent needs and donation offers for each hour. Due to Twitter TOS, we do not share full tweets content on HDX. Please contact us via HDX or on aidr.qcri@gmail.com to get tweet ids of the dataset along with a tool which can be used to rehydrate tweets from tweet ids.

Trainingset Conversion

## Overview

Trainingset Conversion is a dataset for object detection tasks - it contains People annotations for 7,459 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).

Object Detection From Drone

## Overview

Object Detection From Drone is a dataset for object detection tasks - it contains People Cars annotations for 6,423 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.

Here are a few use cases for this project:

1. Social Media Content Categorization - The model can be used in various social media platforms to automatically categorize images based on the content. For example, if an image contains a person, the platform may categorize it under 'People' or 'Portraits', making it easier for users to find specific types of content.

2. Advanced Security Surveillance - The model can be integrated into security systems to identify individuals in surveillance footage. This would improve security measures by allowing for accurate and quick recognition of people.

3. Health and Safety Compliance - For companies needed to ensure social distancing or count the number of people in a facility at a given time, the model could analyze CCTV footage in real-time to measure compliance.

4. Smart Photo Album Management - For personal users, the model can be used in organizing digital photo albums. By identifying the people, pictures can be automatically sorted into specific folders or albums, making it easier for users to navigate their saved images.

5. Autonomous Vehicles - The model could be integrated into the vision systems of autonomous vehicles to help detect and identify people. This would enhance pedestrian detection capabilities, making the vehicles safer.

This version (V3) fixes a bug in Version 2 where 1993 data did not properly deal with missing values, leading to enormous counts of crime being reported. This is a collection of Offenses Known and Clearances By Arrest data from 1960 to 2016. The monthly zip files contain one data file per year(57 total, 1960-2016) as well as a codebook for each year. These files have been read into R using the ASCII and setup files from ICPSR (or from the FBI for 2016 data) using the package asciiSetupReader. The end of the zip folder's name says what data type (R, SPSS, SAS, Microsoft Excel CSV, feather, Stata) the data is in. Due to file size limits on open ICPSR, not all file types were included for all the data. The files are lightly cleaned. What this means specifically is that column names and value labels are standardized. In the original data column names were different between years (e.g. the December burglaries cleared column is "DEC_TOT_CLR_BRGLRY_TOT" in 1975 and "DEC_TOT_CLR_BURG_TOTAL" in 1977). The data here have standardized columns so you can compare between years and combine years together. The same thing is done for values inside of columns. For example, the state column gave state names in some years, abbreviations in others. For the code uses to clean and read the data, please see my GitHub file here. https://github.com/jacobkap/crime_data/blob/master/R_code/offenses_known.RThe zip files labeled "yearly" contain yearly data rather than monthly. These also contain far fewer descriptive columns about the agencies in an attempt to decrease file size. Each zip folder contains two files: a data file in whatever format you choose and a codebook. The data file is aggregated yearly and has already combined every year 1960-2016. For the code I used to do this, see here https://github.com/jacobkap/crime_data/blob/master/R_code/yearly_offenses_known.R.If you find any mistakes in the data or have any suggestions, please email me at jkkaplan6@gmail.comAs a description of what UCR Offenses Known and Clearances By Arrest data contains, the following is copied from ICPSR's 2015 page for the data.The Uniform Crime Reporting Program Data: Offenses Known and Clearances By Arrest dataset is a compilation of offenses reported to law enforcement agencies in the United States. Due to the vast number of categories of crime committed in the United States, the FBI has limited the type of crimes included in this compilation to those crimes which people are most likely to report to police and those crimes which occur frequently enough to be analyzed across time. Crimes included are criminal homicide, forcible rape, robbery, aggravated assault, burglary, larceny-theft, and motor vehicle theft. Much information about these crimes is provided in this dataset. The number of times an offense has been reported, the number of reported offenses that have been cleared by arrests, and the number of cleared offenses which involved offenders under the age of 18 are the major items of information collected.

QDEX_People

## Overview

QDEX_People is a dataset for object detection tasks - it contains Person annotations for 236 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).

Here are a few use cases for this project:

1. E-commerce Inventory Management: The funcaptcha model can be used in e-commerce platforms to automatically categorize products uploaded by sellers based on the objects recognized in the product images. This can significantly improve the efficiency of inventory management and product searches.

2. Trash Sorting App: An app that uses funcaptcha to help users sort their trash. By taking a picture of an item, the model could identify what the item is and tell the user how and where to dispose of it properly.

3. Home Inventory Management: Users can take pictures of their belongings, and the model can identify and catalog them. This could be useful for insurance purposes, moving, or general organization.

4. Educational Game: Developing an educational app for kids in which they can take pictures of various objects, and the app will identify what the object is, helping them learn new words and objects.

5. Assisting Visually Impaired People: funcaptcha can be used in an app that identifies objects in the environment and provides auditory feedback to assist visually impaired users in understanding their surroundings.

Started out as a pumpkin detector to test training YOLOv5. Now suffering from extensive feature creep and probably ending up as a cat/dog/spider/pumpkin/randomobjects-detector. Or as a desaster.

The dataset does not fit https://docs.ultralytics.com/tutorials/training-tips-best-results/ well. There are no background images and the labeling is often only partial. Especially in the humans and pumpkin category where there are often lots of objects in one photo people apparently (and understandably) got bored and did not labe everything. And of course the images from the cat-category don't have the humans in it labeled since they come from a cat-identification model which ignored humans. It will need a lot of time to fixt that.

Dataset used: - Cat and Dog Data: Cat / Dog Tutorial NVIDIA Jetson https://github.com/dusty-nv/jetson-inference/blob/master/docs/pytorch-cat-dog.md © 2016-2019 NVIDIA according to bottom of linked page - Spider Data: Kaggle Animal 10 image set https://www.kaggle.com/datasets/alessiocorrado99/animals10 Animal pictures of 10 different categories taken from google images Kaggle project licensed GPL 2 - Pumpkin Data: Kaggle "Vegetable Images" https://www.researchgate.net/publication/352846889_DCNN-Based_Vegetable_Image_Classification_Using_Transfer_Learning_A_Comparative_Study https://www.kaggle.com/datasets/misrakahmed/vegetable-image-dataset Kaggle project licensed CC BY-SA 4.0 - Some pumpkin images manually copied from google image search - https://universe.roboflow.com/chess-project/chess-sample-rzbmc Provided by a Roboflow user License: CC BY 4.0 - https://universe.roboflow.com/steve-pamer-cvmbg/pumpkins-gfjw5 Provided by a Roboflow user License: CC BY 4.0 - https://universe.roboflow.com/nbduy/pumpkin-ryavl Provided by a Roboflow user License: CC BY 4.0 - https://universe.roboflow.com/homeworktest-wbx8v/cat_test-1x0bl/dataset/2 - https://universe.roboflow.com/220616nishikura/catdetector - https://universe.roboflow.com/atoany/cats-s4d4i/dataset/2 - https://universe.roboflow.com/personal-vruc2/agricultured-ioth22 - https://universe.roboflow.com/sreyoshiworkspace-radu9/pet_detection - https://universe.roboflow.com/artyom-hystt/my-dogs-lcpqe - license: Public Domain url: https://universe.roboflow.com/dolazy7-gmail-com-3vj05/sweetpumpkin/dataset/2 - https://universe.roboflow.com/tristram-dacayan/social-distancing-g4pbu - https://universe.roboflow.com/fyp-3edkl/social-distancing-2ygx5 License MIT - Spiders: https://universe.roboflow.com/lucas-lins-souza/animals-train-yruka

Currently I can't guarantee it's all correctly licenced. Checks are in progress. Inform me if you see one of your pictures and want it to be removed!

Here are a few use cases for this project:

1. Language Learning Assistance: With the "Names" model, users can more easily learn to identify and differentiate between various word classes in the given characters set, improving their reading and pronunciation skills in the languages that use these characters.

2. Optical Character Recognition (OCR): This model can be applied to develop an OCR system for accurately detecting text and word classes in images or scanned documents, aiding transcription, data extraction, and digitization of printed materials using these characters.

3. Speech-to-Text Conversion: The "Names" model can be integrated into speech-to-text systems that handle multiple languages using the given characters set to help accurately transcribe spoken words and phrases, taking into account the identified word classes.

4. Document Analysis and Information Retrieval: Implement the model for analyzing and categorizing documents based on the identified word classes, helping to improve search results, content organization, and knowledge extraction from documents containing these characters.

5. Assistive Technologies: Utilize the "Names" model to develop tools for people with visual impairments, reading difficulties or learning disabilities, enabling them to understand and process text in languages that use the given character set more effectively.