School of Science and Technology

Implementing an Object Detection algorithm in a mobile app

Motivation and scope

Object detection is the task of localizing and classifying objects in an image. Recent scientific progress in deep learning has resulted in a number of powerful machine learning algorithms for object detection. However, these algorithms are often very large and requires a high computational power, such as a GPU. Therefore, many devices that uses object detection often upload the picture to a cloud service that performs the object detection and then returns the results to the device. This solution introduces a delay and limits the range of possible applications.

Specific tasks 

In this project, the task is to implement an object detection algorithm directly in an android or iOS mobile device. The objects that needs to be detected ranges from easier objects such as large vehicles to more challenging smaller objects such as bees.

Necessary skills

Good programming skills in Python. You will use the latest mobile app support from Tensorflow lite or PyTorch 1.3. Basic knowledge in machine learning.


You will learn how to train an object detection algorithm and implement in a mobile device.


Martin Längkvist