Isaac's Blog

While many deep learning-based methods have been developed to enable 3D instance segmentation, such as SGPN and ASIS, it is still interesting to see how some traditional methods can perform on this task, for example, using 3D shape contexts. The shape context is a kind of regional point descripto...

Class imbalance is a common problem in machine learning, where the total number of one class of data is far less than that of another class of data. Class imbalance affects the quality and reliability of results in machine learning tasks as most of the evaluation metrics assume a balanced class d...

We travel just to travel. The world looks different when you bike. A car window is more or less a divide between the traveler and the outside world, a motorcycle engine could be too noisy for a person who loves a peaceful trip. Biking, however, makes you taste the true beauty of nature: sufferi...

Plane detection is a widely used technique that can be applied in many applications, e.g., augmented reality (AR), where we have to detect a plane to generate AR models, and 3D scene reconstruction, especially for man-made scenes, which consist of many planar objects. Nowadays, with the prolifera...

Being a novel deep net architecture invariant towards input order, PointNet is able to consume unordered point clouds directly and thus has a promising prospect in the field of geometry processing. At present, the most popular implementation of PointNet is based on TensorFlow and it takes HDF5 as...

The RAPter gives us a very nice design and implementation for doing man-made scene reconstruction. According to the nature of its algorithm, it also shows a potential use in terms of 3D plane detection. However, the documentation related to it is pretty poor, so I decided to write here a summary ...

In geometry processing, Tutte embedding, also known as barycentric embedding, can be used for mesh parameterization by fixing the boundary vertices of the mesh on a certain convex polygon, and building a crossing-free straight-line embedding with the interior vertices inside the convex boundary. ...

The Hough transform is a method for detecting parameterized objects, typically used for lines and circles in 2D space. Nowadays, with the proliferation of acquisitive devices, deriving a massive point cloud is an easy task. As we can also parameterize objects in 3D, Hough transform can be applied...

Mesh smoothing, also known as mesh denoising, is an important and widely discussed topic in geometry processing. Basically, a mesh smoothing method takes three steps: loading a mesh from a file; smoothing that mesh; outputting the mesh to a file. Recently, I implemented a few basic algorithms reg...

The Voronoi diagram and the Delaunay triangulation are dual representations of a set of points to each other. Due to their wide applications in science and technology, the Voronoi diagram and the Delaunay triangulation play important roles in the field of computational geometry. In this post, I a...