COMPUTER VISION LAB
Edge Detection
Available
Feature Extraction
Available
Object Detection
Available
Image Segmentation
Available
Pose Estimation
Available
Neural Radiance Fields
Render Available
Our computer vision research lab has successfully implemented several cutting-edge techniques for image analysis, including edge detection, feature extraction, object detection with YOLO, and image segmentation with SAM2.
Select a technique from above to explore demos and benchmarks, or visit the home page to try these techniques on generated images.
Recent Implementations
Image Segmentation with SAM2
We've implemented Segment Anything Model 2 (SAM2) for high-quality image segmentation. The implementation features a pink background for better visualization and uses center-point prompting to identify the main object in the image.
Pose Detection with TensorFlow
Our human pose detection system uses a TensorFlow model to identify body keypoints and draw a skeleton overlay. The system can detect 18 different body parts and connect them with lines to visualize the human pose.
Object Detection with YOLOv3
Our object detection system uses YOLOv3 to identify and classify objects in images with adjustable confidence thresholds.
Neural Radiance Fields (NeRF)
Our NeRF implementation creates detailed 3D models from 2D images, allowing for novel view synthesis. Check out our truck demonstration to see how we've applied this technology to create interactive 3D visualizations from video footage.
