Video Analysis For Broadcasting, Retail & Security.

Vehicle Detection

Detects different types of vehicles and pedestrians in a video stream or still image and returns the classification and the object’s position.


  1. Accurate vehicle counting, real-time traffic analysis, speed measurement, hazard detection.
  2. Real-time parking occupation.
  3. Lane occupation, violations detection.

Automatic License Number Recognition

Detects vehicle license plate numbers. It recognises the vehicle type and color and compares it with the plate number when data are available.


  1. Counting the type of vehicles based on pollutant emissions and correlation with environmental data
  2. Verify access permissions in both public and private areas (car parks, Restricted Traffic Zones)
  3. Check insurance coverage and stolen vehicles

License Plate Recognition – Blurred To Protect Privacy

Face Detection

Detects and identifies human faces within an image and returns the positions. This is the first step for subsequent processing.


  1. Accurate counting of the number of people visible in a photo or video frame (counting people passing and/or watching a shop window or monitor)
  2. Extraction of face images for later processing

Face Recognition

Recognise a specific person whose images are already available (one ID photo is generally enough). It can be used for user verification and logging, and for searching specific people within video.

Watch a continuous flow of people and find out what people are passing again under the camera.


  1. Seek specific people within a crowd
  2. Detect abnormal behaviour

Face Similarity

Find faces similar to a reference face, from a given collection of faces, returns a collection of similar faces and relative metrics of similarity.


  1. Group similar faces into large photo libraries
  2. React to the images of visitors to a showcase or store based on their morphological characteristics (gender, age, etc.)

Face Detection And Recognition

Face Keypoints

Find the key points of the facial components (face contour, eye, eyebrow, lip and nose contour).


  1. “Virtual” make-up, punctual variations (eye color change for example), variations for Identikit (beard insertion, glasses, etc.).
  2. Avatar generation, enhanced reality

Face Morphological Characteristics

Measures visual attributes including Age, Gender, Ethnicity, Smile Intensity, Head Position, Eye Direction.


  1. Direct Advertising (shows what’s matter to the specific customer)
  2. Market research, customer feedback measurement (count who’s passing or looking to a shop window)

Face Keypoints And Morphological Measures

People Detection (Full Body)

Detects and locates multiple people within cluttered environments and returns position, direction and speed.


  1. Monitoring the number of people, people density and crowd flow. Detecting unusual activities such as assemblages, blocks, and queues for both commercial and public security purposes.
  2. Counting people for commercial and security reasons.
  3. Automatic camera alignment for close-up shots

Automatic People Detection And People Counting (Full Body) – Cropped

Objects And Luggage Detection (Security Applications)

Detects and locates specific objects (like trolleys and backpacks) in cluttered video streams, returns position, direction and speed.


  1. Detecting unusual activities
  2. Find people carrying specific objects