Diksha Shukla

Mohamad Zamini, Diksha Shukla

Open-vocabulary semantic segmentation requires assigning pixel-level semantic labels while supporting an open and unrestricted set of categories. Training-free CLIP-based approaches preserve strong zero-shot generalization but typically rely on a single inference mechanism, limiting their ability to jointly address unreliable local tokens and insufficient spatial coherence. We propose DouC, a training-free dual-branch CLIP framework that decomposes dense prediction into two complementary components. OG-CLIP improves patch-level reliability via lightweight, inference-time token gating, while FADE-CLIP injects external structural priors through proxy attention guided by frozen vision foundation models. The two branches are fused at the logit level, enabling local token reliability and structure-aware patch interactions to jointly influence final predictions, with optional instance-aware correction applied as post-processing. DouC introduces no additional learnable parameters, requires no retraining, and preserves CLIP's zero-shot generalization. Extensive experiments across eight benchmarks and multiple CLIP backbones demonstrate that DouC consistently outperforms prior training-free methods and scales favorably with model capacity.

Mohamad Zamini, Diksha Shukla

Multimodal Large Language Models (MLLMs) combine visual and textual representations to enable rich reasoning capabilities. However, the high computational cost of processing dense visual tokens remains a major bottleneck. A critical component in this pipeline is the visual projector, which bridges the vision encoder and the language model. Standard designs often employ a simple multi-layer perceptron for direct token mapping, but this approach scales poorly with high-resolution inputs, introducing significant redundancy. We present Delta-LLaVA, a token-efficient projector that employs a low-rank DeltaProjection to align multi-level vision features into a compact subspace before further interaction. On top of this base alignment, lightweight Transformer blocks act as specialization layers, capturing both global and local structure under constrained token budgets. Extensive experiments and ablations demonstrate that this base-then-specialize design yields consistent gains across multiple benchmarks with only 144 tokens, highlighting the importance of token formation prior to scaling interaction capacity. With Delta-LLaVA, inference throughput improves by up to 55%, while end-to-end training accelerates by nearly 4-5x in pretraining and over 1.5x in finetuning, highlighting the dual benefits of our design in both efficiency and scalability.

Rajesh Kumar, Partha Pratim Kundu, Diksha Shukla et al.

In this paper, we propose a novel continuous authentication system for smartphone users. The proposed system entirely relies on unlabeled phone movement patterns collected through smartphone accelerometer. The data was collected in a completely unconstrained environment over five to twelve days. The contexts of phone usage were identified using k-means clustering. Multiple profiles, one for each context, were created for every user. Five machine learning algorithms were employed for classification of genuine and impostors. The performance of the system was evaluated over a diverse population of 57 users. The mean equal error rates achieved by Logistic Regression, Neural Network, kNN, SVM, and Random Forest were 13.7%, 13.5%, 12.1%, 10.7%, and 5.6% respectively. A series of statistical tests were conducted to compare the performance of the classifiers. The suitability of the proposed system for different types of users was also investigated using the failure to enroll policy.