This dataset contain video transcript from a limited number of youtubers who post Their review on iPhone 15, 15 plus , pro and pro max model . These are the videos used for the videos. Video Credits are owned by respective creators.

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F13244501%2Fc3bf6524f3ddfa376794de29f97651a1%2F_results_14_0.png?generation=1695205189424943&alt=media" alt="">

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F13244501%2F645638973f5f8f5782cc8720ac4214c1%2F_results_15_0.png?generation=1695205202162850&alt=media" alt="">

For more check Here

Apple is one of the most influential and recognisable brands in the world, responsible for the rise of the smartphone with the iPhone. Valued at over $2 trillion in 2021, it is also the most valuable...

The number of smartphone users in the United States was forecast to continuously increase between 2024 and 2029 by in total 17.4 million users (+5.61 percent). After the fifteenth consecutive increasing year, the smartphone user base is estimated to reach 327.54 million users and therefore a new peak in 2029. Notably, the number of smartphone users of was continuously increasing over the past years.Smartphone users here are limited to internet users of any age using a smartphone. The shown figures have been derived from survey data that has been processed to estimate missing demographics.The shown data are an excerpt of Statista's Key Market Indicators (KMI). The KMI are a collection of primary and secondary indicators on the macro-economic, demographic and technological environment in up to 150 countries and regions worldwide. All indicators are sourced from international and national statistical offices, trade associations and the trade press and they are processed to generate comparable data sets (see supplementary notes under details for more information).Find more key insights for the number of smartphone users in countries like Mexico and Canada.

The number of Apple iPhone unit sales dramatically increased between 2007 and 2023. Indeed, in 2007, when the iPhone was first introduced, Apple shipped around *** million smartphones. By 2023, this number reached over *** million units. The newest models and iPhone’s lasting popularity Apple has ventured into its 17th smartphone generation with its Phone ** lineup, which, released in September 2023, includes the **, ** Plus, ** Pro and Pro Max. Powered by the A16 bionic chip and running on iOS **, these models present improved displays, cameras, and functionalities. On the one hand, such features come, however, with hefty price tags, namely, an average of ***** U.S. dollars. On the other hand, they contribute to making Apple among the leading smartphone vendors worldwide, along with Samsung and Xiaomi. In the first quarter of 2024, Samsung shipped over ** million smartphones, while Apple recorded shipments of roughly ** million units. Success of Apple’s other products Apart from the iPhone, which is Apple’s most profitable product, Apple is also the inventor of other heavy-weight players in the consumer electronics market. The Mac computer and the iPad, like the iPhone, are both pioneers in their respective markets and have helped popularize the use of PCs and tablets. The iPad is especially successful, having remained as the largest vendor in the tablet market ever since its debut. The hottest new Apple gadget is undoubtedly the Apple Watch, which is a line of smartwatches that has fitness tracking capabilities and can be integrated via iOS with other Apple products and services. The Apple Watch has also been staying ahead of other smart watch vendors since its initial release and secures around ** percent of the market share as of the latest quarter.

In the first quarter of its 2025 fiscal year, Apple generated around ** billion U.S. dollars in revenue from the sales of iPhones. Apple iPhone revenue The Apple iPhone is one of the biggest success stories in the smartphone industry. Since its introduction to the market in 2007, Apple has sold more than *** billion units worldwide. As of the third quarter of 2024, the Apple iPhone’s market share of new smartphone sales was over ** percent. Much of its accomplishments can be attributed to Apple’s ability to keep the product competitive throughout the years, with new releases and updates. Apple iPhone growth The iPhone has shown to be a crucial product for Apple, considering that the iPhone’s share of the company’s total revenue has consistently grown over the years. In the first quarter of 2009, the iPhone sales were responsible for about ********* of Apple’s revenue. In the third quarter of FY 2024, this figure reached a high of roughly ** percent, equating to less than ** billion U.S. dollars in that quarter. In terms of units sold, Apple went from around **** million units in 2010 to about *** million in 2023, but registered a peak in the fourth quarter of 2020 with more than ** million iPhones sold worldwide.

We are publishing a walking activity dataset including inertial and positioning information from 19 volunteers, including reference distance measured using a trundle wheel. The dataset includes a total of 96.7 Km walked by the volunteers, split into 203 separate tracks. The trundle wheel is of two types: it is either an analogue trundle wheel, which provides the total amount of meters walked in a single track, or it is a sensorized trundle wheel, which measures every revolution of the wheel, therefore recording a continuous incremental distance.
Each track has data from the accelerometer and gyroscope embedded in the phones, location information from the Global Navigation Satellite System (GNSS), and the step count obtained by the device. The dataset can be used to implement walking distance estimation algorithms and to explore data quality in the context of walking activity and physical capacity tests, fitness, and pedestrian navigation. Methods The proposed dataset is a collection of walks where participants used their own smartphones to capture inertial and positioning information. The participants involved in the data collection come from two sites. The first site is the Oxford University Hospitals NHS Foundation Trust, United Kingdom, where 10 participants (7 affected by cardiovascular diseases and 3 healthy individuals) performed unsupervised 6MWTs in an outdoor environment of their choice (ethical approval obtained by the UK National Health Service Health Research Authority protocol reference numbers: 17/WM/0355). All participants involved provided informed consent. The second site is at Malm ̈o University, in Sweden, where a group of 9 healthy researchers collected data. This dataset can be used by researchers to develop distance estimation algorithms and how data quality impacts the estimation.

All walks were performed by holding a smartphone in one hand, with an app collecting inertial data, the GNSS signal, and the step counting. On the other free hand, participants held a trundle wheel to obtain the ground truth distance. Two different trundle wheels were used: an analogue trundle wheel that allowed the registration of a total single value of walked distance, and a sensorized trundle wheel which collected timestamps and distance at every 1-meter revolution, resulting in continuous incremental distance information. The latter configuration is innovative and allows the use of temporal windows of the IMU data as input to machine learning algorithms to estimate walked distance. In the case of data collected by researchers, if the walks were done simultaneously and at a close distance from each other, only one person used the trundle wheel, and the reference distance was associated with all walks that were collected at the same time.The walked paths are of variable length, duration, and shape. Participants were instructed to walk paths of increasing curvature, from straight to rounded. Irregular paths are particularly useful in determining limitations in the accuracy of walked distance algorithms. Two smartphone applications were developed for collecting the information of interest from the participants' devices, both available for Android and iOS operating systems. The first is a web-application that retrieves inertial data (acceleration, rotation rate, orientation) while connecting to the sensorized trundle wheel to record incremental reference distance [1]. The second app is the Timed Walk app [2], which guides the user in performing a walking test by signalling when to start and when to stop the walk while collecting both inertial and positioning data. All participants in the UK used the Timed Walk app.

The data collected during the walk is from the Inertial Measurement Unit (IMU) of the phone and, when available, the Global Navigation Satellite System (GNSS). In addition, the step count information is retrieved by the sensors embedded in each participant’s smartphone. With the dataset, we provide a descriptive table with the characteristics of each recording, including brand and model of the smartphone, duration, reference total distance, types of signals included and additionally scoring some relevant parameters related to the quality of the various signals. The path curvature is one of the most relevant parameters. Previous literature from our team, in fact, confirmed the negative impact of curved-shaped paths with the use of multiple distance estimation algorithms [3]. We visually inspected the walked paths and clustered them in three groups, a) straight path, i.e. no turns wider than 90 degrees, b) gently curved path, i.e. between one and five turns wider than 90 degrees, and c) curved path, i.e. more than five turns wider than 90 degrees. Other features relevant to the quality of collected signals are the total amount of time above a threshold (0.05s and 6s) where, respectively, inertial and GNSS data were missing due to technical issues or due to the app going in the background thus losing access to the sensors, sampling frequency of different data streams, average walking speed and the smartphone position. The start of each walk is set as 0 ms, thus not reporting time-related information. Walks locations collected in the UK are anonymized using the following approach: the first position is fixed to a central location of the city of Oxford (latitude: 51.7520, longitude: -1.2577) and all other positions are reassigned by applying a translation along the longitudinal and latitudinal axes which maintains the original distance and angle between samples. This way, the exact geographical location is lost, but the path shape and distances between samples are maintained. The difference between consecutive points “as the crow flies” and path curvature was numerically and visually inspected to obtain the same results as the original walks. Computations were made possible by using the Haversine Python library.

Multiple datasets are available regarding walking activity recognition among other daily living tasks. However, few studies are published with datasets that focus on the distance for both indoor and outdoor environments and that provide relevant ground truth information for it. Yan et al. [4] introduced an inertial walking dataset within indoor scenarios using a smartphone placed in 4 positions (on the leg, in a bag, in the hand, and on the body) by six healthy participants. The reference measurement used in this study is a Visual Odometry System embedded in a smartphone that has to be worn at the chest level, using a strap to hold it. While interesting and detailed, this dataset lacks GNSS data, which is likely to be used in outdoor scenarios, and the reference used for localization also suffers from accuracy issues, especially outdoors. Vezovcnik et al. [5] analysed estimation models for step length and provided an open-source dataset for a total of 22 km of only inertial walking data from 15 healthy adults. While relevant, their dataset focuses on steps rather than total distance and was acquired on a treadmill, which limits the validity in real-world scenarios. Kang et al. [6] proposed a way to estimate travelled distance by using an Android app that uses outdoor walking patterns to match them in indoor contexts for each participant. They collect data outdoors by including both inertial and positioning information and they use average values of speed obtained by the GPS data as reference labels. Afterwards, they use deep learning models to estimate walked distance obtaining high performances. Their results share that 3% to 11% of the data for each participant was discarded due to low quality. Unfortunately, the name of the used app is not reported and the paper does not mention if the dataset can be made available.

This dataset is heterogeneous under multiple aspects. It includes a majority of healthy participants, therefore, it is not possible to generalize the outcomes from this dataset to all walking styles or physical conditions. The dataset is heterogeneous also from a technical perspective, given the difference in devices, acquired data, and used smartphone apps (i.e. some tests lack IMU or GNSS, sampling frequency in iPhone was particularly low). We suggest selecting the appropriate track based on desired characteristics to obtain reliable and consistent outcomes.

This dataset allows researchers to develop algorithms to compute walked distance and to explore data quality and reliability in the context of the walking activity. This dataset was initiated to investigate the digitalization of the 6MWT, however, the collected information can also be useful for other physical capacity tests that involve walking (distance- or duration-based), or for other purposes such as fitness, and pedestrian navigation.

The article related to this dataset will be published in the proceedings of the IEEE MetroXRAINE 2024 conference, held in St. Albans, UK, 21-23 October.

This research is partially funded by the Swedish Knowledge Foundation and the Internet of Things and People research center through the Synergy project Intelligent and Trustworthy IoT Systems.

Infant Crying Smartphone speech dataset, collected by Android smartphone and iPhone, covering infant crying. Our dataset was collected from extensive and diversify speakers(201 people in total, with balanced gender distribution), geographicly speaking, enhancing model performance in real and complex tasks. Quality tested by various AI companies. We strictly adhere to data protection regulations and privacy standards, ensuring the maintenance of user privacy and legal rights throughout the data collection, storage, and usage processes, our datasets are all GDPR, CCPA, PIPL complied.

This dataset can be used for Sentiment Analysis which contains the tweets about apple products on twitter. This data set has basically 3 headers 1. tweet_text 2.emotion_in_tweet_is_directed_at 3.is_there_an_emotion_directed_at_a_brand_or_product

What’s Included

Introducing the English Product Image Dataset - a diverse and comprehensive collection of images meticulously curated to propel the advancement of text recognition and optical character recognition (OCR) models designed specifically for the English language.

Containing a total of 2000 images, this English OCR dataset offers diverse distribution across different types of front images of Products. In this dataset, you'll find a variety of text that includes product names, taglines, logos, company names, addresses, product content, etc. Images in this dataset showcase distinct fonts, writing formats, colors, designs, and layouts.

To ensure the diversity of the dataset and to build a robust text recognition model we allow limited (less than five) unique images from a single resource. Stringent measures have been taken to exclude any personally identifiable information (PII) and to ensure that in each image a minimum of 80% of space contains visible English text.

Images have been captured under varying lighting conditions – both day and night – along with different capture angles and backgrounds, to build a balanced OCR dataset. The collection features images in portrait and landscape modes.

All these images were captured by native English people to ensure the text quality, avoid toxic content and PII text. We used the latest iOS and Android mobile devices above 5MP cameras to click all these images to maintain the image quality. In this training dataset images are available in both JPEG and HEIC formats.

Along with the image data, you will also receive detailed structured metadata in CSV format. For each image, it includes metadata like image orientation, county, language, and device information. Each image is properly renamed corresponding to the metadata.

The metadata serves as a valuable tool for understanding and characterizing the data, facilitating informed decision-making in the development of English text recognition models.

We're committed to expanding this dataset by continuously adding more images with the assistance of our native English crowd community.

If you require a custom product image OCR dataset tailored to your guidelines or specific device distribution, feel free to contact us. We're equipped to curate specialized data to meet your unique needs.

Furthermore, we can annotate or label the images with bounding box or transcribe the text in the image to align with your specific project requirements using our crowd community.

This Image dataset, created by FutureBeeAI, is now available for commercial use.

Leverage the power of this product image OCR dataset to elevate the training and performance of text recognition, text detection, and optical character recognition models within the realm of the English language. Your journey to enhanced language understanding and processing starts here.

Description

SWAN-DF: the first high fidelity publicly available dataset of realistic audio-visual deepfakes, where both faces and voices appear and sound like the target person. The SWAN-DF dataset is based on the public SWAN database of real videos recorded in HD on iPhone and iPad Pro (in year 2019). For 30 pairs of manually selected people from SWAN, we swapped faces and voices using several autoencoder-based face swapping models and using several blending techniques from the well-known open source repo DeepFaceLab and voice conversion (or voice cloning) methods, including zero-shot YourTTS, DiffVC, HiFiVC, and several models from FreeVC.

For each model and each blending technique, there are 960 video deepfakes. We used three types of models of the following resolutions: 160x160, 256x256, and 320x320 pixels. We took one pre-trained model corresponding for each resolution, and tuned it for each of the 30 pairs (both ways) of subjects for 50K iterations. Then, when generating deepfake videos for each pair of subjects, we used one of the tuned models and a way to blend the generated image back into the original frame, which we call blending technique. SWAN-DF dataset contains 25 different combinations of models and blending, which means the total number of deepfake videos is 960*25=24000.

We generated speech deepfakes using four voice conversion methods: YourTTS, HiFiVC, DiffVC, and FreeVC. We did not use text to speech methods for our video deepfakes, since the speech they produce is not synchronized with the lip movements in the video. For YourTTS, HiFiVC, and DiffVC methods, we used the pretrained models provided by the authors. HiFiVC was pretrained on VCTK, DiffVC on LibriTTS, and YourTTS on both VCTK and LibriTTS datasets. For FreeVC, we generated audio deepfakes for several variants: using the provided pretrained models (for 16Hz with and without pretrained speaker encoder and for 24Hz with pretrained speaker encoder) as is and by tuning 16Hz model either from scratch or starting from the pretrained version for different number of iterations on the mixture of VCTK and SWAN data. In total, SWAN-DF contains 12 different variations of audio deepfakes: one for each of YourTTS, HiFiVC, and DiffVC and 9 variants of FreeVC.

Acknowledgements

If you use this database, please cite the following publication:

Pavel Korshunov, Haolin Chen, Philip N. Garner, and Sébastien Marcel, "Vulnerability of Automatic Identity Recognition to Audio-Visual Deepfakes", IEEE International Joint Conference on Biometrics (IJCB), September 2023. https://publications.idiap.ch/publications/show/5092

About This Dataset

The Roboflow Thermal Dogs and People dataset is a collection of 203 thermal infrared images captured at various distances from people and dogs in a park and near a home. Some images are deliberately unannotated as they do not contain a person or dog (see the Dataset Health Check for more). Images were captured both portrait and landscape. (Roboflow auto-orient assures the annotations align regardless of the image orientation.)

Thermal images were captured using the Seek Compact XR Extra Range Thermal Imaging Camera for iPhone. The selected color palette is Spectra.

Example

This is an example image and annotation from the dataset: https://i.imgur.com/h9vhrqB.png" alt="Man and Dog">

Usage

Thermal images have a wide array of applications: monitoring machine performance, seeing in low light conditions, and adding another dimension to standard RGB scenarios. Infrared imaging is useful in security, wildlife detection,and hunting / outdoors recreation.

This dataset serves as a way to experiment with infrared images in Roboflow. (Or, you could build your own night time pet finder!)

Collecting Custom Data

Roboflow is happy to improve your operations with infrared imaging and computer vision. Services range from data collection to building automated monitoring systems leveraging computer vision. Reach out for more.

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:

The dataset created at the 2024 National Festival in Portugal consists of 9062 images. It is a dataset designed for recognizing objects from the YCB dataset along with detecting people. It comprises 1500 images of people from the COCO dataset, 416 images taken in the arena with the Astra camera of the TIAGo robot, 2146 images taken with an iPhone, and the rest of the images are from the @Home Objects Dataset.

Object Tracking on a Monopoly Game Board

Author: Nathan Hoebeke

Supervisors: Maxim Van de Wynckel, Prof. Dr. Beat Signer

About

The goal of this dataset was to track game pieces on the physical game board of Monopoly. We make use of object classification where our training data consists of 100 pictures (taken at an angle) of the game board in order to classify the individual (moving) pieces. The training dataset was on the 9th of April 2023 and the test date recorded on the 7th of May 2023 using an iPhone 13 mini and iPhone 12.

Two participants played a game of Monopoly and each individually took pictures of the current game state after every move. These images were then processed by our application to determine the location of pawns and other game pieces such as the red and green houses.

Raw images are unprocessed but may have minor edits to ensure anonymisation of participants in the background. We used Roboflow to label and train our dataset which is included in this repository.

For more information about our processing and this dataset you can download the full Bachelor thesis here: https://wise.vub.ac.be/thesis/location-tracking-physical-game-board (download link available after embargo at the end of the academic year)

This dataset was published as part of the bachelor thesis: Location Tracking on a Physical Game Board for obtaining the degree of Bachelor in Computer Sciences at the Vrije Universiteit Brussel.

Data

Dataset contents

• model: Trained YOLOv5 model with labels. This dataset can also be found here.
• train: Training data made by the author.
  • raw: Raw pictures of the game board at various states.
    • GAME_2023-04-09_: Images formatted based on the date and time when they were captured.
  • processed: Processed pictures with perspective transformation applied.
    • canvas: (Pre)-processed image.
• test: Test data made by indepedent participants.
  • participant_1: Participant 1 data
    • raw: Raw pictures of the game board taken by the parcipant after every move.
      • GAME_2023-05-07_: Images formatted based on the date and time when they were captured.
    • processed: Processed pictures with perspective transformation applied. Yellow rectangles are included when our own algorithm was able to determine the location.
      • canvas: Processed image.
  • participant_2: Participant 2 data
    • raw: Raw pictures of the game board taken by the parcipant after every move.
      • GAME_2023-05-07_: Images formatted based on the date and time when they were captured.
• README.md: Documentation and information about the dataset.

License

This license applies to the dataset for the game Monopoly. Any artwork or intellectual property from the game that is captured by this dataset is property of Hasbro, Inc.

Copyright 2022-2023 Nathan Hoebeke, Beat Signer, Maxim Van de Wynckel, Vrije Universiteit Brussel

Permission is hereby granted, free of charge, to any person obtaining a copy of this dataset and associated documentation files (the “Dataset”), to deal in the Dataset without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Dataset, and to permit persons to whom the Dataset is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions that make use of the Dataset.

THE DATASET IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE DATASET OR THE USE OR OTHER DEALINGS IN THE DATASET.

The Ink and Identity dataset presents a comprehensive collection of 374 handwritten characters consisting of all lowercase and 12 uppercase alphabets, including a few digits, gathered from participants spanning diverse age groups. Utilized A4-size unruled paper for writing. As of now, the datasets available are on ruled paper containing only a single line or single word, which limits analyzing data in graphology, and all the datasets do not contain all alphabets, like numbers and characters. These limitations are covered in this dataset.

The dataset was collected using two methods: an in-person (300 images) and a web form (150 images). Total dataset: 450 images. The dataset comprises two folders, in-person and web form, and the image was captured in a 1:1 ratio on an iPhone 14 Plus 12MP. Web form images were instructed to be captured in a 1:1 ratio, and each one has been pre-processed with a unique ID like “001” and stored in .jpg format.

Additionally, researchers can employ this dataset as a reference standard for age-based handwriting analysis, personality trait prediction, AI/ML model training, cognitive and neurological research, and health and mental well-being.

Face Antispoofing dataset for recognition systems

The dataset consists of 98,000 videos and selfies from 170 countries, providing a foundation for developing robust security systems and facial recognition algorithms.

While the dataset itself doesn't contain spoofing attacks, it's a valuable resource for testing liveness detection system, allowing researchers to simulate attacks and evaluate how effectively their systems can distinguish between real faces and various forms of spoofing.

By utilizing this dataset, researchers can contribute to the development of advanced security solutions, enabling the safe and reliable use of biometric technologies for authentication and verification. - Get the data

Examples of data

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F22059654%2Fe46e401a5449bacce5f934aaea9bb06e%2FFrame%20155.png?generation=1730591437955112&alt=media" alt=""> The dataset offers a high-quality collection of videos and photos, including selfies taken with a range of popular smartphones, like iPhone, Xiaomi, Samsung, and more. The videos showcase individuals turning their heads in various directions, providing a natural range of movements for liveness detection training.

💵 Buy the Dataset: This is a limited preview of the data. To access the full dataset, please contact us at https://unidata.pro to discuss your requirements and pricing options.

Metadata for the dataset

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F22059654%2F8350718e93ee92840995405815739c61%2FFrame%20136%20(1).png?generation=1730591760432249&alt=media" alt=""> Furthermore, the dataset provides detailed metadata for each set, including information like gender, age, ethnicity, video resolution, duration, and frames per second. This rich metadata provides crucial context for analysis and model development.

Researchers can develop more accurate liveness detection algorithms, which is crucial for achieving the iBeta Level 2 certification, a benchmark for robust and reliable biometric systems that prevent fraud.

🌐 UniData provides high-quality datasets, content moderation, data collection and annotation for your AI/ML projects

iBeta Level 1 - liveness detection dataset

The dataset was collected specifically for iBeta requirements for Level 1 certification of the Liveness Detection system.

💴 For Commercial Usage: Full version of the dataset includes 42,000 video attacks, leave a request on TrainingData to buy the dataset

The dataset consists of more than 42,000 video attacks of 7 different types specifically curated for evaluating liveness detection algorithms. The dataset aims to incorporate different scenarios and challenges to enable robust assessment and comparison of liveness detection systems.

The iBeta Liveness Detection Level 1 dataset serves as a benchmark for the development and assessment of liveness detection systems and for evaluating and improving the performance of algorithms.

Types of attacks:

1. Outline: printed outlines of people's photos.
2. Outline3d: printed portraits of people attached to a cylinder.
3. Mask: printed portraits of people with cut-out eyes in various variations.
4. Mask3d: portraits consisting of several connected cardboard masks.
5. Phone: demonstrations of a person's photo on the smartphone.
6. Monitor: demonstrations of a person's photo on a computer screen.
7. Real: a real human face

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F12421376%2F72c514066335ed6b6d1d273d1c4198ef%2FMASK_TYPES.png?generation=1678913343424357&alt=media" alt="">

Each attack was filmed on an Apple iPhone and Google Pixel. The videos were filmed on various backgrounds and using additional accessories such as faked facial hair, scarves, hats, and others.

The dataset comprises videos of genuine facial presentations using various methods, including 3D masks and photos, as well as real and spoof faces. It proposes a novel approach that learns and extracts facial features to prevent spoofing attacks, based on deep neural networks and advanced biometric techniques.

Our results show that this technology works effectively in securing most applications and prevents unauthorized access by distinguishing between genuine and spoofed inputs. Additionally, it addresses the challenging task of identifying unseen spoofing cues, making it one of the most effective techniques in the field of anti-spoofing research.

💴 Buy the Dataset: This is just an example of the data. Leave a request on https://trainingdata.pro/datasets to discuss your requirements, learn about the price and buy the dataset

iBeta Level 2 Liveness Detection

Try the iBeta Level 2 dataset: Link to the Kaggle

TrainingData provides high-quality data annotation tailored to your needs.

keywords: ibeta level 1, ibeta level 2, liveness detection systems, liveness detection dataset, biometric dataset, biometric data dataset, replay attack dataset, biometric system attacks, anti-spoofing dataset, face liveness detection, deep learning dataset, face spoofing database, face anti-spoofing, presentation attack detection, presentation attack dataset, 2D print attacks, 3D print attacks, phone attack dataset, face anti spoofing, large-scale face anti spoofing, rich annotations anti spoofing dataset, cut prints spoof attack

Videos:4,349,945 (Users:2,060,573) 3,573,499 of videos are 3+ seconds

Gender: - Male 60% - Female 40%

Ethnicity: - Asian 9% - African Decent 13% - East Indian 3% - Latino Hispanic 28% - Caucasian 47%

Age Group: - 0-17 3% - 18-24 62% - 25-34 21% - 35-44 10% - 45-54 3% - 55+ 1%

Top Phone Models: - iPhone 6s 9% - iPhone XR 6% - iPhone 6 6% - iPhone 7 (US/CDMA) 6% - iPhone 11 5% - iPhone 8 (US/CDMA) 4%

Top Countries: - US 48.84% - GB 10.57% - CA 4.26% - AU 3.48% - FR 2.80% - SA 2.17%

Includes 474 000 users absent from the selfie photo set; the rest typically have one selfie and ~2 short voiced videos.

All videos are collected with the consent of users.

10 People - 3D&2D Living_Face & Anti_Spoofing Data. The collection scenes is indoor scenes. The dataset includes males and females. The age distribution ranges from juvenile to the elderly, the young people and the middle aged are the majorities. The device includes iPhone X, iPhone XR. The data diversity includes various expressions, facial postures, anti-spoofing samples, multiple light conditions, multiple scenes. This data can be used for tasks such as 3D face recognition, 3D Living_Face & Anti_Spoofing.

Data from: Maps made with smartphones highlight lower noise pollution during COVID-19 pandemic lockdown at four locations in Boston

https://doi.org/10.5061/dryad.ncjsxkt35

Dataset contents include csv files of all data (each file describes collection year and site of data), R script used to create noise maps, and kml files needed to run the map creation code.

Description of the data and file structure

Each csv file contains the L50 values (median sound level) taken from hundreds of 20 second recordings over multiple collection days. The SPLnFFT application exports the latitude and longitude of where the recording was taken, which is also included in the csv files and is used to create the noise maps. The csv files are used as data frames for the R script to create noise maps for each collection site. The R script contains comments and instructions to clearly indicate each step of the map creation. The kml files are used to create bound...

Total Users 10,229,822 Total Pictures 10M+ (mostly 1 per ID)

Gender: - Male 60% - Female 40%

Ethnicity: - Asian 9% - African Decent 13% - East Indian 3% - Latino Hispanic 28% - Caucasian 47%

Age Group: - 0-17 3% - 18-24 62% - 25-34 21% - 35-44 10% - 45-54 3% - 55+ 1%

Top Phone Models: - iPhone 6s 9% - iPhone XR 6% - iPhone 6 6% - iPhone 7 (US/CDMA) 6% - iPhone 11 5% - iPhone 8 (US/CDMA) 4% (Total 141 device)

Top Countries: - US 48.84% - GB 10.57% - CA 4.26% - AU 3.48% - FR 2.80% - SA 2.17% (Total 131 countries)

Average resolution 5761024 px.

All photos are collected with the consent of users.