Pattern Recognition Application

Amphibian Identification App


Pattern Recognition System

Type: Android mobile app
Technologies: Android, Java, Kotlin, OpenCV, PostgreSQL, Spring boot 2, Spring Data, Vaadin 10
Duration: 5 months
Team 1 project manager, 2 analysts, 4 developers and 1 QA specialist,


A US based client asked us to build an app that can identify progressive changes within the population of amphibian (like frogs and newts) life cycle.

The amphibian lifecycle is more complicated compared to other animals, since they undergo metamorphosis.
assignment_turned_in

Problem:

The existing methods of tracking amphibians is not efficient, is costly and can adversely affect amphibian’s well being and health conditions.
An app was built using computer vision and machine learning technology to accurately identify amphibians in the wild by unique pattern recognition matched against large number of species in its database.

Challenge

Required to create a machine learning algorithm to accurately identify amphibians directly from their captured images/photos taken by user scientists using their smartphone camera.

Solution

Our AI engineers did a fantastic job to design the algorithm logic, but we also had to create a smart workflow logic as well. This endeavor was accomplished by our UI/UX engineers and statistical program analyst to build a user friendly interphase for smooth delivery during their research.

Results

The finished app is used as a tool by nature scientists to complement their work of research.

They are able to capture photos of the subject amphibian, identify it, add description and store it under the right classification in the database.

Thanks to the great work of our team of engineers designing this computer vision algorithm, scientists can seamlessly view the numeric values of captured images of each subject amphibian in the database for further population analysis.

Areas of our expertise

Machine Learning
Google Maps
Camera API
Computer Vision
Custom UI Widgets
Image Processing
User Setting Persistence

Other technologies

Dagger (dependency Injection)
Retrofit
Ktor