---
title: "Facial Keypoint Detection Dataset"
description: "5,000+ photos"
url: "https://unidata.pro/datasets/facial-keypoint-detection/"
date_modified: "2025-10-11T19:02:09+03:00"
language: "en-US"
---
Large-scale face dataset for facial keypoint detection containing nearly 200,000 annotated images of diverse individuals displaying facial emotions across various age groups and scenes. Designed for training facial keypoint detection, emotion recognition, expression analysis, and facial recognition models using rich facial features and metadata to support advanced computer vision applications.

🤖 [View as Markdown](https://unidata.pro/datasets/facial-keypoint-detection.md)

## Dataset Structure

### The Numbers Section

**Numbered list:**

- **Number:** 5,000+ — **Text:** photos
- **Number:** 15 — **Text:** landmarks

### Tooltips Section

**Tooltip items:**

- **Name:** Data annotation
- **Name:** Computer Vision
- **Name:** Facial Recognition
- **Name:** Anti-spoofing
- **Name:** Machine learning

### Dataset Information

**Table with data:**

| Characteristic | Data |
| --- | --- |
| Description | close-up images of faces with annotated data for facial recognition |
| Data types | Image |
| Tasks | Facial recognition, Computer Vision |
| Number of images | 5,000+ |
| Number of landmarks | 15 |
| Landmarks | Left eye, the closest point to the nose, Left eye, pupil's center, Left eye, the closest point to the left ear, Right eye, the closest point to the nose, Right eye, pupil's center, Right eye, the closest point to the right ear, Left eyebrow, the closest point to the nose, Left eyebrow, the closest point to the left ear, Right eyebrow, the closest point to the nose, Right eyebrow, the closest point to the right ear, Nose, center, Mouth, left corner point, Mouth, right corner point, Mouth, the highest point in the middle, Mouth, the lowest point in the middle |
| Labeling | Metadata (ID, gender) |
| Gender | Male, Female |

**Media Slider:**

- **Image in the slider:** ![Original image](https://unidata.pro/wp-content/uploads/2024/12/facial-keypoint-detection-original-image.webp)
- **Image in the slider:** ![Labeling of the image](https://unidata.pro/wp-content/uploads/2024/12/facial-keypoint-detection-labeling-of-the-image.webp)

**Link to the sample:** [Download sample](https://drive.google.com/drive/u/0/folders/1WY8XQ5QS3eyBmUOv1mKxDeyOKSLocOwc)

### Technical Specifications

**Table with data:**

| Characteristic | Data |
| --- | --- |
| Image extension | JPG |
| Extension of labeling file | JSON |

**Source and data collection methodology:** Source and collection methodology. Data was collected by UniData team by using the crowdsourcing service

### Dataset Use Cases - Slider

**Industry Cards:**

- **Industry:** Computer Vision & Machine Learning — **Title:** Training Models for Facial Landmark Detection — **Text:** Facial Keypoint Detection Dataset provides detailed facial keypoints and annotated landmarks on human faces, making it ideal training data for developing keypoints detection and facial recognition models. The datasets contain accurately labeled keypoint positions, enabling algorithms to learn detection methods that improve detection accuracy and model robustness in real-world conditions.
- **Industry:** Human-Computer Interaction & Emotion Analysis — **Title:** Recognizing Facial Expressions and Emotions — **Text:** This facial detection dataset supports research in expression recognition and emotion recognition by mapping facial landmarks that correspond to different facial expressions. The dataset consists of diverse human faces displaying various emotional states, helping developers train recognition algorithms that interpret facial expressions for applications like virtual assistants, interactive avatars, and psychological behavior studies.
- **Industry:** Augmented Reality & 3D Modeling — **Title:** Improving Face and Pose Estimation Systems — **Text:** Facial Keypoints Detection Dataset is valuable for building 3D faces and pose estimation systems. By providing labeled head poses, key points, and bounding boxes, the dataset helps train face detector models used in AR filters, motion tracking, and object detection frameworks that rely on precise landmark detection.
- **Industry:** Security & Biometric Identification — **Title:** Enhancing Accuracy in Face Recognition Systems — **Text:** This keypoint dataset strengthens facial recognition and biometric verification solutions by improving landmark detection and keypoint detector performance. The datasets consist of carefully annotated custom datasets with validation and test splits, enabling training processes that refine recognition algorithms for secure authentication, surveillance, and identity verification applications in modern face recognition technologies.

### Fact

**FAQs Heading:** FAQs

**List of Questions:**

- **Question:** How was the dataset collected? — **Answer:** The dataset was created through crowdsourced data collection, where contributors provided high-quality facial images under different lighting and pose conditions. This ensures that the dataset covers a wide range of human faces, head poses, and facial expressions for robust model training.
- **Question:** What are the sources of data for Unidata datasets? — **Answer:** Facial Keypoint Detection Dataset was collected by the Unidata team using a crowdsourcing service. All images feature real human faces captured under various lighting and pose conditions to provide diverse training data for facial recognition systems.
- **Question:** Can I request a sample of the dataset before purchasing? — **Answer:** Yes. Unidata provides free dataset samples for evaluation purposes. This allows you to review the image quality, annotation accuracy, and facial landmark labeling before purchasing the full version for facial keypoints detection or emotion recognition projects.
- **Question:** Is it possible to request a custom dataset? — **Answer:** Yes. Unidata offers custom datasets designed to meet your specific needs. You can request additional facial landmarks, different demographic distributions, or annotation formats to support your recognition algorithms and machine learning experiments.
- **Question:** How are Unidata datasets licensed? — **Answer:** Unidata datasets follow a dual-licensing model. Free samples are available for testing and validation, while the full dataset is available for purchase for research, training, or commercial use.
- **Question:** Do Unidata datasets follow GDPR or other data privacy regulations? — **Answer:** Yes. All datasets are developed in full compliance with GDPR and relevant data protection laws. The data is collected ethically through legal channels, ensuring the responsible handling of personal and biometric information.
- **Question:** How are Unidata datasets stored? — **Answer:** Unidata securely stores all datasets on AWS cloud infrastructure. This ensures data protection, high availability, and scalability, following ISO 27001 and ISO 27701 standards to maintain international information security and privacy management compliance.
- **Question:** Why are facial landmarks important for machine learning? — **Answer:** Facial landmarks provide precise reference points that allow AI models to understand the structure and movement of a face. Accurate landmark localization improves downstream tasks such as facial recognition, emotion analysis, face tracking, and biometric verification.

[Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)