MD Gulam Hussain

Academic Research Trainee

Led research and development of ACAF-Net, a dual-stream RGB-Thermal fusion network for real-time edge surveillance, achieving state-of-the-art results on the FLIR ADAS v2 dataset. Optimized model for edge GPUs using quantization-aware training and structured sparsity, reducing FLOPs by 51.5% and achieving 53 FPS real-time INT8 inference.

PyTorchRepViTVision Transformers (ViT)Quantization Aware Training (QAT)TensorRTComputer VisionEdge AI