This project uses YOLOv2 for human detection and stereo cameras for distance measurement. It runs on a PYNQ-Z2 board with a neural network accelerator on the PL side and outputs video via HDMI.

Built a convolutional neural network on the PYNQ-Z2 platform and accelerated traffic sign recognition using FPGA.

SAT-MTB-VSR dataset and codes for "A Lightweight Recurrent Aggregation Network for Satellite Video Super-Resolution", Journal of Selected Topics in Applied Earth Observations and Remote Sensing (JS…

[ICCV 2025] Cross-modal Ship Re-Identification via Optical and SAR Imagery: A Novel Dataset and Method

The custom modifitions in original mmdetection workspace for making it can be competent with multi modal instance segmentation.