Résumé‎ > ‎

Available Master's Projects

1) HDL Implementation of Motion Estimation

The sum of absolute differences (SAD) algorithm measures the similarity between image blocks. It works by taking the absolute difference between each pixel in the original block and the corresponding pixel in the block being used for comparison. These differences are summed to create a simple metric of block similarity, the L1 norm of the difference image or Manhattan distance between two image blocks. SAD may be used for a variety of purposes, such as object recognition, the generation of disparity maps for stereo images, and motion estimation for video compression. SAD is also easily parallelizable since it analyzes each pixel separately, making it practical to implement on an FPGA.

2) Fast Image Encryption by Scrambling of JPEG Quantization Table

Image security is important for Internet commerce. However, traditional cryptographic algorithms such as AES are not fast enough to process the vast amount of image data generated by multimedia applications. The main idea of this project is to incorporate cryptographic techniques with digital image processing techniques to achieve fast image encryption and decryption while providing a considerable level of security.

3) GPU Implementation of JPEG Image Compression

The Raspberry Pi is a single-board computer which features an ARM11 processor and a graphics processing unit (GPU) optimized for for parallel computing of data at relatively low clock speeds.  The objective of this project is to implement JPEG compression on the Raspberry Pi GPU. Some core algorithms which can be accelerated include Discrete Cosine Transform (DCT), Huffman encoding and run-length encoding.

4) GPU Implementation of JPEG Domain Motion Detection

The Raspberry Pi is a single-board computer which features an ARM11 processor and a graphics processing unit (GPU) optimized for for parallel computing of data at relatively low clock speeds. The objective of this project is to compare inter-frame blocks in the JPEG domain without full decompressing the JPEG encoded picture. This will save considerable amounts of unnecessary decompression.  The frames compared in JPEG domain will be benchmarked against fully-decompressed frames.

5) Bird Species Identification by Frequency Analysis of Bioacoustics

Identification of bird species residing in a particular locale is an important step in biodiversity and conservation. Automated species identification will enable remote sites to be monitored. Bird sounds recorded in the location of interest will processed using speech-recognition based algorithms to identify the bird species. The database of bird sounds is based on the website xeno-canto.org.

Teaching  Research  Publications  Consultancy  Projects  Skills