We leverage wavelets to decompose images into different spatial resolutions across multiple frequency bands, enabling precise modifications across different levels of detail.

Undergraduate thesis project. Improved YOLOv4-tiny for real-time object detection on Raspberry Pi.

Designed a head-mounted smart glasses system for circuit inspection with real-time communication and defect detection capabilities.

Proposed a lightweight object detection architecture to balance accuracy and inference speed for edge devices.

Introduced a novel loss function and attention mechanism to improve unsupervised domain adaptation performance.

Proposed a wavelet-based Score Distillation Sampling (SDS) method to generate high-quality textures for complex mesh objects.

Resolving multi-view inconsistency in 3D generation by leveraging geometry-conditioned diffusion models.

Investigated ML tools for pre/post-compensation of linear and nonlinear signal distortions in optical transmitters.