Advanced Topics in Digital Image Processing


Provide the students with advanced knowledge of digital image processing and GPU processing.

General characterization





Responsible teacher

André Teixeira Bento Damas Mora


Weekly - 4

Total - 56

Teaching language



It is recommended that the students attend the course of Sensorial Systems before taking TAPDI, in order to learn the basics of digital image processing.


Gonzalez, R., & Woods, R. (2007). Digital Image Processing - third edition. Prentice-Wall.

Munshi, A., Gaster, B., Mattson, T. G., & Ginsburg, D. (2011). OpenCL Programming Guide (p. 648). Pearson Education. Retrieved from

Scarpino, M. (2011). OpenCL in Action: How to Accelerate Graphics and Computation (p. 434). Manning Publications Company.

Teaching method

Theoretical classes for presentation of the basic concepts and practical classes for the development of small exercises that explore the underlying concepts. The practical classes will be based on guidelines previously provided to guide the students in the preparation of the practical exercises. The course final project will seek to integrate the largest number of concepts in order to provide an opportunity for integration and consolidation of knowledge.

Evaluation method

Students will be evaluated in the theoretical component through a written research paper, which will be presented publicly during classes and a set of moodle tests. In the practical component they will develop a final project, which will be discussed with the teachers of the discipline.

NF = 55% * ( 60% * N_article + 20% * N_presentation + 20% * T_Moodle ) + 45% * N_Pratical

Subject matter

Image compression techniques (Huffman coding and JPEG compression)
- Discrete cosine transform

Frequency domain image enhancement
- 2D Fourier Transform

Image restoration or refocusing
- Wiener Filtering

Image Segmentation
- Watershed Transform
- Gradient Path Labeling

Hough and Radon Transforms

Pattern Recognition
- Template Matching
- Viola-Jones Algorithm
- Histogram of Oriented Gradients (HOG)


High Speed Image

Thermal Imaging

GPU Image Processing - OpenCL
- OpenCL Architecture
- Optimization