Academic
Academic
Home
Experience
Projects
Publications
CV
Light
Dark
Automatic
Deep Learning
A Jacobi-set based loss function for segmentation task
Segmentation of fine-scale structures in natural and bio-medical images are gaining importance with the development of high resolution electron microscopy images. The task still remains challenging as per-pixel accuracy is not only the metric of concern because of the imbalance in the dataset. In this project, a new loss function based on the Jacobi-sets are proposed.
Soham Mukherjee
PDF
Graph generation with Geometrical and Topological Constraints
Persistent homology, a tool from computational topology we computed persistent diagrams of graphs and incorporated topological & geometrical constraints while generating graphs.
GEFL: Extended Filtration Learning for Graph Classification
Learning Extended filtration on graphs
Simon Zhang
,
Soham Mukherjee
,
Tamal K. Dey
PDF
Code
GEFL: Extended Filtration Learning for Graph Classification
Extended persistence is a technique from topological data analysis to obtain global multiscale topological information from a graph. This includes information about connected components and cycles that are captured by the so-called persistence barcodes. We introduce extended persistence into a supervised learning framework for graph classification. Global topological information, in the form of a barcode with four different types of bars and their explicit cycle representatives, is combined into the model by the readout function which is computed by extended persistence. The entire model is end-to-end differentiable
Simon Zhang
,
Soham Mukherjee
,
Tamal K. Dey
PDF
Code
Cite
×