Embedded Files
Enhancing 3D Object Detection for cars by introducing a novel approach to Frustum-PointPillars architecture using YOLO and PSP-Net segmentation model leveraging multi-stage sensor fusion of RGB and LiDAR data.
Enhancing 3D Object Detection for cars by introducing a novel approach to Frustum-PointPillars architecture using YOLO and PSP-Net segmentation model leveraging multi-stage sensor fusion of RGB and LiDAR data.
Computer Vision / Deep Learning/ Autonomous Driving
Computer Vision / Deep Learning/ Autonomous Driving
Experimented with Vision Transformer (ViT) against adversarial patches to build robust models, yielding 63% robust accuracy on ImageNet.
Experimented with Vision Transformer (ViT) against adversarial patches to build robust models, yielding 63% robust accuracy on ImageNet.
Computer Vision / Deep Learning/ Autonomous Driving
Computer Vision / Deep Learning/ Autonomous Driving
Developed an unsupervised protoNN (based on K-Nearest Neighbours) model to detect Freezing of Gait (FoG) and achieved a 95% detection accuracy with a 2KB model size.
Developed an unsupervised protoNN (based on K-Nearest Neighbours) model to detect Freezing of Gait (FoG) and achieved a 95% detection accuracy with a 2KB model size.
Data Analysis / Machine Learning
Data Analysis / Machine Learning
Page updated
Google Sites
Report abuse