Welcome to the showcase of my deep learning and computer vision projects, where innovation meets practical application. My portfolio spans a variety of cutting-edge projects that demonstrate my expertise in harnessing the power of neural networks to interpret and understand visual data. From developing sophisticated algorithms for real-time object detection and recognition, to crafting advanced neural network architectures for image segmentation and enhancement, each project embodies my commitment to advancing the field of computer vision.

GANs

LiDAR SLAM

1. Segmentation using Superpixels
2. Generative Adversarial Networks(GANs) on MNIST and Deep Convolutional GANs(DCGANs) on Human Faces Dataset
3. Simultaneous Localization and Mapping using LiDAR data

Semantic Segmentation and Object Detection

1. Sample Segmentation Results
2. Sample Object Detection Results

Implicit Neural Representation

An illustration of an image represented in the implicit neural representation (INR) form, i.e. as an MLP that predicts an RGB pixel value given its (x, y) coordinates. In contrast to a pixel-based representation, it stores the image in its true continuous form.

Structure from Motion

Pose Estimation

Classical Computer Vision

K-Means Clustering
Stereo Depth Estimation
Lane detection
Trajectory Tracking
Zhang’s Camera Calibration
Tsai Camera Calibration
Image Stitching
Camera Pose Estimation

Deep Learning/CV

Classical Computer Vision

References