The Challenge in Person Re-ID Technology For Smart City

Nowadays, face recognition technology has been widely used in the development of smart city. In surveillance video, due to the camera resolution and shooting angle, it is usually impossible to capture very high-quality face pictures. In addition, one camera cannot cover all areas, and there is often no overlap between multiple cameras.

When face recognition fails, Re-ID becomes a very important complementary technology. Re-ID realizes cross-camera tracking of pedestrians by using the overall pedestrian features as an important supplement to the face.

Person Re-ID, full name Person Re-Identification, uses computer vision technology to detect whether there is a specific pedestrian in an image or video. Re-ID can re-identify the same person in uncertain scenes through clothing, posture, hairstyle, etc., and describe the individual’s travel trajectory with these features. This technology is widely used in intelligent video surveillance and intelligent security.

Althrough Re-ID technology has made a lot of progress, it still faces many practical problems and technical difficulties. These problems mainly include the difficulty of data acquisition and algorithm training.

● Data Acquisition

Compared with face data, Re-ID data is seriously scarce. The most mainstream dataset has only 1000–3000 pedestrian IDs, while the size of the public face dataset has exceeded 1 million. The main reason for this phenomenon is that the pedestrian dataset needs to be collected from the same person under multiple cameras at the same time over a period of time, which restricts the construction of the Re-ID dataset.

● Algorithm Training

Based on the shortage of training data, the existing video surveillance is restricted by factors such as imaging quality and resolution, resulting in unclear captured image information. Of course, there are also factors such as large differences in camera shooting angles, changes in indoor and outdoor environments, changes in pedestrian clothing and accessories, large differences in seasonal clothing styles, and differences in light during the day and night. These factors make Re-ID analysis more difficult.

In order to help ReID technology solve the above problems, Datatang has developed the ReID dataset. Datatang’s ReID dataset includes the data collected by 21,000 people in real scenarios and controlled construction scenarios.

10,000 People Real Scene Re-ID Data

The data includes 10,000 collectors in real scenes such as shopping malls, supermarkets, and communities. Each scene has an average of about 15 cameras, covering a variety of monitoring heights, monitoring shooting angles, and monitoring areas (for example, the same shopping mall has different monitoring areas) The human body information, and there are occlusion truncations that occur in real scenes.

10,000 People Controlled Scene Re-ID Data

In order to solve the identification difficulty of the same person changing different clothes and different people wearing the same clothes, the data was collected in a controlled scene, and the data hall built the collection scene by itself to form a 360-degree full-angle monitoring, a total of 12 cameras, one camera every 30 degrees.

1033 People Monitoring Scene Data

In order to increase the richness of human body poses, a total of 1033 people were collected in this dataset, and each subject collected 30 different poses. At the same time, in order to increase the diversity of angles, each subject collects ReID data from head-up and top-down views.

Datatang’s ReID dataset far exceeds open source resources in terms of the scale of the collected people and the number of cross-cameras, and covers a variety of scenes. In addition, Datatang’s ReID datasets have been authorized by the collectors, strictly complying with ISO27701 privacy management system and ISO27001 information security management system and customers can use them with confidence.


If you want to know more details about the datasets or how to acquire, please feel free to contact us:

Off-the-shelf AI training data, on-demand data collection & annotation services

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Explore the Technical Approaches to Mitigating Algorithmic Bias

To get the reproduction number down to 1 as the new strain becomes the dominant one

Non-negative Matrix Factorization (NMF)

Enriching OCR with Akka Streams

Keeping Up With Data — Week 19 Reading List

Bulk Import Makes Multiple Patient Registration Smooth

Why Clean Data is Vital for Nonprofits Using Salesforce

Forecasting Store Traffic Using Feature Selection with sklearn and Pandas in Python

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store


Off-the-shelf AI training data, on-demand data collection & annotation services

More from Medium

Performing Scenario Identification and Categorization, a Crucial Step in AD/ADAS Development

Adelta Tutorial — Part 1: Differentiating a Simple Shader Program

Deep Learning for Human Activity Recognition

Frequently Asked Questions Regarding Computer Vision