Master
2022/2023
Project seminar "Computer vision for mobile devises"
Type:
Compulsory course (Master of Computer Vision)
Area of studies:
Applied Mathematics and Informatics
Delivered by:
Department of Applied Mathematics and Informatics
When:
2 year, 3 module
Mode of studies:
offline
Open to:
students of one campus
Master’s programme:
Master of Computer Vision
Language:
English
ECTS credits:
7
Contact hours:
3
Course Syllabus
Abstract
Students of the online course will acquire practical skills for development of computer vision algorithms on Android devices. We implement traditional algorithms and neural networks for filtering and matching images, searching for keypoints, classifying and detecting objects. Students study how to use OpenCV library, TensorFlow and PyTorch frameworks to recognize scenes on photos from a gallery, implement facial analytics, transfer style of famous painting into their photos, etc.
Learning Objectives
- The objective of this course is learning the details of implementation of both traditional image processing methods and neural network models for Android mobile applications.
- The main expected practical learning outcome is mastering programming skills of image processing on mobile devices.
Expected Learning Outcomes
- Distinguish differences between C++ and Java programming
- Set-up environment for Android development
- Display video from frontal and rear cameras of mobile device
- Use NDK for embedded programming with C++
- Apply OpenCV on mobile platforms for image enhancements
- Perform image classification on Android using pre-trained neural networks
- Compress models for efficient image processing on mobile devices
- Use PyTorch and TfLite on Android device
- Find correspondent parts in different images
- Create panoramas with image stitching on mobile device
- Detect and recognize text on Android
- Implement object detection and scene segmentation on mobile devices
- Use neural style transfer to transform photos in a gallery
- Implement face detection using either OpenCV or deep neural networks
- Recognize age, gender and emotions on mobile devices
Course Contents
- Overview of programming for Android
- Basic image processing operations on mobile devices
- Image classification with deep neural networks in mobile applications
- Image matching on mobile devices
- Applied image processing tasks for mobile devices
- Mobile facial analytics
Bibliography
Recommended Core Bibliography
- Sebastian Raschka, & Vahid Mirjalili. (2019). Python Machine Learning : Machine Learning and Deep Learning with Python, Scikit-learn, and TensorFlow 2, 3rd Edition. Packt Publishing.
- Машинное обучение и TensorFlow : пер. с англ., Шакла, Н., Фриклас, К., 2019
Recommended Additional Bibliography
- Прикладное машинное обучение с помощью Scikit-Learn и TensorFlow : концепции, инструменты и техники для создания интеллектуальных систем: пер. с англ., Жерон, О., 2018