Master's Thesis - Uppsala, Sverige - Vidhance AB

Beskrivning

The master's thesis projects are a very important part of exploratory work at Vidhance. Every year we invite a number of students to perform master's thesis projects in areas we find to be of interest for future opportunities. Many of our previous thesis students have also started working with us after finishing their master's thesis.

Here you will find the thesis topics we have available for 2023, and you're welcome to apply using the "apply for this job" button. Please specify which topic or topics you are interested in when writing your application, and attach a transcript of records.

1. Machine Learning-based Approaches for Video Saliency Detection in Video Clips
2. Real-time Neural Network Performance Optimization for Video Stream Processing on Modern Smartphones
3. Automated and Anonymous Large-Scale Data Collection of Camera-Related Information from Mobile Devices

Machine Learning-based Approaches for Video Saliency Detection in Video Clips

This master's thesis project aims to explore and develop novel machine learning-based methods for video saliency detection in video clips. The core objective of the research is to advance the understanding and implementation of video saliency detection mechanisms, which could be instrumental in areas like video editing, video compression, video surveillance, augmented reality, and virtual reality.

Background:

Saliency detection is an essential component of the video analysis process, aiming to identify the most visually striking parts of a video frame. Traditional methods often rely on handcrafted features and pre-determined algorithms, which may not fully grasp the complexity of video content in diverse scenarios. As machine learning has demonstrated promising results in many areas of computer vision, it offers potential improvements in saliency detection.

Real-time Neural Network Performance Optimization for Video Stream Processing on Modern Smartphones

The primary objective of this master's thesis project is to investigate and develop methods for optimizing the real-time performance of neural networks used for processing video streams on modern smartphones. With an ever-increasing number of mobile applications utilizing machine learning algorithms to handle live video data, the efficient and effective operation of these neural networks on mobile devices presents a significant challenge.

Background:

Even with consistent improvements in mobile hardware, there are limitations in processing power, memory, and energy that create challenges for deploying sophisticated neural networks on mobile devices. Furthermore, the real-time processing of live video streams adds another layer of complexity due to the substantial computational requirements and the need for rapid, reliable results.

Automated and Anonymous Large-Scale Data Collection of Camera-Related Information from Mobile Devices

This master's thesis project aims to investigate tools for automated and anonymous large-scale data collection from mobile devices, such as smartphones, specifically focusing on camera-related data. The purpose is to explore the available methods for collecting anonymous yet useful data and to compare various solutions and their pros and cons. The project includes implementing and evaluating the proposed techniques to collect and analyze the gathered data.

Objectives:

Understand the available techniques for collecting anonymous and useful data from mobile devices.
* Focus on camera-related data points, such as the number of videos, video lengths, exposure times, focus settings, white balance, scene types, shakiness, lighting conditions, location, the number of people in the videos, connection speed, device information (brand, model, version, OS), time of capture, FoV, resolution, and OIS.
* Implement and compare different data collection solutions and draw conclusions about their advantages and disadvantages.

Background:

With the growing usage of smartphones, valuable data can be collected from these devices to improve user experience and optimize various applications. Collecting anonymous data ensures the users' privacy, while providing useful insights for further research and development in the field of mobile devices and camera-related technologies.

If you have questions regarding any of these projects, we're happy to answer them. Please connect with us on this career portal of ours to start a conversation.