DASIP 2023: Workshop on Design and Architectures for Signal and Image Processing

10:00 – 10:10 DASIP’23 Welcome and Opening

10:10 – 11:00 Keynote #1: Real-time 4K video stream processing on SoC FPGAs.
Tomasz Kryjak, AGH University of Science and Technology, Poland
Abstract:

The keynote will summarise the results of a 5-year research project on the hardware implementation of algorithms for processing 4K resolution video streams in real-time on SoC FPGAs (System on Chip Field Programmable Gate Arrays). First, the specifics of the 4K signal and the challenges of the so-called vector format (2 or 4 pixels per clock) will be presented. This will be followed by a discussion of the implemented modules and algorithms – from very basic point and context operations, to methods such as foreground object segmentation, optical flow, CLAHE (Contrast Limited Adaptive Histogram Equalization) filtering,  ORB (Oriented FAST and Rotated BRIEF) feature point detection and matching, HOG+SVM (Histogram of Oriented Gradients + Support Vector Machine) pedestrian detection, object tracking (Siamese networks, correlation filters), connected component labeling and stereovision. In the final part of the presentation, the current research of the Embedded Vision Systems Team based at AGH University of Science and Technology in Krakow will be presented.

11:00 – 11:30 Coffee Break

11:30 – 13:00 Session #1: Methods and Applications I
Chair: Tiago M. Dias, ISEL – Técnico Lisboa

• 11:30 – 11:55 SCAPE: HW-Aware Clustering of Dataflow Actors for Tunable Scheduling Complexity.
  Ophélie Renaud, Dylan Gageot, Karol Desnos and Jean-François Nezan.
• 11:55 – 12:20 Deep Recurrent Neural Network performing spectral recurrence on hyperspectral images for brain tissue classification.
  Pedro L. Cebrián, Alberto Martín-Pérez, Manuel Villa, Jaime Sancho, Gonzalo Rosa, Guillermo Vazquez, Pallab Sutradhar, Alejandro Martinez de Ternero, Miguel Chavarrías, Alfonso Lagares, Eduardo Juarez and Cesar Sanz.
• 12:20 – 12.45 Brain blood vessel segmentation in hyperspectral images through linear operators.
  Guillermo Vázquez, Manuel Villa, Alberto Martín-Pérez, Jaime Sancho, Gonzalo Rosa, Pedro L. Cebrián, Pallab Sutradhar, Alejandro Martinez de Ternero, Miguel Chavarrías, Alfonso Lagares, Eduardo Juarez and Cesar Sanz.

13:00 – 14:00 Lunch

14:00 – 15.30 Session #2: Hardware Architectures
Chair: Alfonso Rodríguez, Universidad Politécnica de Madrid, Spain

• 14:00 – 14:25 AINoC: new interconnect for future Deep Neural Network accelerators.
  Hana Krichene, Rohit Prasad and Ayoub Mouhagir.
• 14:25 – 14:50 Real-time FPGA implementation of the Semi-Global Matching stereo vision algorithm for an 4K/UHD video stream.
  Mariusz Grabowski and Tomasz Kryjak.
• 14:50 – 15:15 TaPaFuzz – An FPGA-Accelerated Framework for RISC-V IoT Graybox Fuzzing.
  Florian Meisel, David Volz, Christoph Spang, Nguyen Tien Dat Tran and Andreas Koch.

15:30 – 16:00 Coffee Break