Deep Learning - one shot learning for speaker recognition using Filter Banks

Signature verification system using Siamese networks

Visual Object Tracking algorithms. Hold on! There is a lot to come

A PyTorch implementation of siamese networks using backbone from torchvision.models, with support for TensorRT inference.

The repository of SiamHAN, an IPv6 address correlation model on TLS encrypted traffic. The work has been accepted as USENIX Security 2021 accepted Paper.

The code for Fisher Discriminant Triplet (FDT) and Fisher Discriminant Contrastive (FDC) loss functions

SiamBOMB: Siamese network using Background information for real-time Online Multi-species home-cage animal Behavioral analysis. (Updating)

The code for offline and online triplet mining for Siamese networks

Official implementation of the NASiam paper.

Face Recognition using Siamese(Twin) Network along with Triplet Loss.

Implementation of state-of-the-art NLP models using transformers for tasks including machine translation, text-summarization, chatbots, and question answering.

Implementation of Siamese network for handwritten Arabic characters using Keras.

Bachelor's graduation project

VOT-Paper-Summary: The objective of this repo is to provide Visual Object Tracking papers' overview concisely.

The project implements Siamese Network with Triplet Loss in Keras to learn meaningful image representations in a lower-dimensional space. By training on the MNIST dataset, it creates a powerful architecture and implements Triplet Loss function. The resulting model enables applications like image search, recommendation systems, and image clustering.

Pytorch implementation of SNN

ActiSiamese (Neurocomputing 2022)

My implementation of Siamese Network for MNIST Dataset in Pytorch and Tensorflow

The Facenet paper of 2015 proposed an interesting solution for huge multiclass problems. Instead of the traditional approach, we try to learn a similarity function i.e. degree of difference between 2 inputs. If the degree of difference between the inputs is less than a threshold then the inputs are classified as similar else different.

Facial comparison using Siamese Networks