Robert McCraith

Can we detect cars in 3D without 3D labels? We present a method for converting pre-trained 2D mask predictions to 3D Object Detections without any supervision on the target dataset. Paper: arxiv.org/abs/2109.07945 Code: github.com/robertmccraith…

Added more models to MIMM (MLX Image Models) Now includes classification (Alexnet, VGG, Resnet, Mobilenetv3, ViT) and detection (DETR) Includes code to convert PyTorch weights and example training + validation github.com/robertmccraith…

To be presented at @ieee_ras_icra session on Computer Vision for Transportation at 16:30 today

Can we detect cars in 3D without 3D labels? We present a method for converting pre-trained 2D mask predictions to 3D Object Detections without any supervision on the target dataset. Paper: arxiv.org/abs/2109.07945 Code: github.com/robertmccraith…