arXiv subjects

Ushnish Sarkar

Publications and source records attributed to Ushnish Sarkar.

4 recordsLinked to original sources

arXiv2026

A HamNoSys-Guided Dataset and Baselines for Fine-Grained Isolated Handshape Recognition in Sign Language

Purpose: Fine-grained handshape recognition supports computational sign-language transcription, recognition, and translation, but broad, phonetically defined visual inventories with signer-aware evaluation remain limited. This work introduces a benchmark grounded in the language-independent Hamburg Notation System (HamNoSys). Methods: A balanced dataset of 144,000 RGB images was collected from 15 participants for 160 handshape classes defined by the official HamNoSys 4 Handshapes Chart. ResNet-18 and ViT-B/16 were evaluated as appearance-based models, while a graph convolutional network and XGBoost were evaluated from hand landmarks. Both a class-stratified subject-dependent split and a 15-fold leave-one-subject-out (LOSO) protocol were used. The same model families were additionally assessed on LSWH100 and ASL Fingerspelling Dataset A for external context. Results: The subject-dependent benchmarks established reproducible reference performance across all four model families, whereas LOSO evaluation exposed a substantial reduction when recognition was required to generalise to unseen participants. On ASL Fingerspelling Dataset A, mean LOSO top-1 accuracy ranged from 82.20% to 87.40%. Conclusion: The documented acquisition, curation, and complementary evaluation protocols pro-vide a reproducible resource for fine-grained isolated-handshape research and for developing more accessible sign-language technologies.

cs.CV↗

arXiv2024

Enhancing ASL Recognition with GCNs and Successive Residual Connections

This study presents a novel approach for enhancing American Sign Language (ASL) recognition using Graph Convolutional Networks (GCNs) integrated with successive residual connections. The method leverages the MediaPipe framework to extract key landmarks from each hand gesture, which are then used to construct graph representations. A robust preprocessing pipeline, including translational and scale normalization techniques, ensures consistency across the dataset. The constructed graphs are fed into a GCN-based neural architecture with residual connections to improve network stability. The architecture achieves state-of-the-art results, demonstrating superior generalization capabilities with a validation accuracy of 99.14%.

cs.CV↗

arXiv2019

On the Universality and Extremality of graphs with a distance constrained colouring

A lambda colouring (or $L(2,1)-$colouring) of a graph is an assignment of non-negative integers (with minimum assignment $0$) to its vertices such that the adjacent vertices must receive integers at least two apart and vertices at distance two must receive distinct integers. The lambda chromatic number (or the $λ$ number) of a graph $G$ is the least positive integer among all the maximum assigned positive integer over all possible lambda colouring of the graph $G$. Here we have primarily shown that every graph with lambda chromatic number $t$ can be embedded in a graph, with lambda chromatic number $t$, which admits a partition of the vertex set into colour classes of equal size. It is further proved that if an $n-$vertex graph with lambda chromatic number $t\geq5$, where $n\geq t+1$, contains maximum number of edges, then the vertex set of such graph admits an equitable partition. For such an admitted equitable partition there are either $0$ or $\min\{|A|,|B|\}$ number of edges between each pair $(A,B)$ of subsets (i.e. roughly, such partition is a "sparse like" equitable partition). Here we establish a classification result, identifying all possible $n-$vertex graphs with lambda chromatic number $t\geq3$, where $n\geq t+1$, which contain maximum number of edges. Such classification provides a solution of a problem posed more than two decades ago by John P. Georges and David W. Mauro.

math.CO↗

arXiv2009

Some Conjectures on the Number of Primes in Certain Intervals

In this paper, we make some conjectures on prime numbers that are sharper than those found in the current literature. First we describe our studies on Legendre's Conjecture which is still unsolved. Next, we show that Brocard's Conjecture can be proved assuming our improved version of Legendre's Conjecture. Finally, we sharpen the Bertrand's Postulate for prime numbers. Our results are backed by extensive empirical investigation.

math.NT↗