Abstract: Generating accurate image captions and retrieving semantically matched image-text pairs are typically treated as independent problems, limiting the depth of visual-semantic understanding a model can develop. This talk presents ViTCap-R, a unified framework that integrates a Vision Transformer encoder, a patch-level attention LSTM decoder, and a dual-encoder contrastive retrieval module into a single cohesive pipeline. By attending over 196 spatial patch embeddings during decoding, the model produces interpretable captions grounded in specific image regions. A shared 256-dimensional embedding space, trained with symmetric InfoNCE loss and hard negative mining, enables effective cross-modal retrieval and caption reranking at inference time. We evaluate ViTCap-R on Flickr8k and MS COCO 2014, reporting BLEU, METEOR, and Recall@K metrics alongside qualitative attention heatmaps, PCA projections, and t-SNE visualizations of the learned embedding space. Results demonstrate that unified training improves semantic alignment between modalities, while patch-level attention offers meaningful interpretability gains over standard global-feature baselines. We also discuss the challenges of scaling to high-diversity datasets and the trade-offs introduced by hard negative mining, pointing toward future directions involving transformer-based decoders and larger-scale pretraining. Speaker(s): Prayash Das Agenda: Hybrid event, in-person or online: 2:30 Introduction Technical Talks Question and Answer Networking 4:30 Conclusion Room: Meeting Rooms 2,3, 2 Civic Center Drive, East Brunswick, New Jersey, United States, 08816, Virtual: https://events.vtools.ieee.org/m/558867

Robotics open house for K-12 Students in Montclair Collaborative Robotics and Smart Systems Laboratory. Room: 448, Bldg: CCIS, 1 Normal Ave, Montclair, New Jersey, United States

Please come and join us for the IEEE North Jersey Section Picnic. This event will be held on Saturday August 1st, 2026 at Duke Island Park in Bridgewater, NJ. You will find as Oak Grove Pavillion. Please come and join us Co-sponsored by: North Jersey Section AP/MTT Chapter Duke Island Park, Oak Grove Pavillion, Bridgewater, New Jersey, United States, 08807

In the world of functional Neuroimaging, the focus has historically been on "Gray Matter" (the brain's processors), often ignoring the "White Matter" (the communication cables). This talk introduces WhiFuN, a pipeline designed to extract meaningful signal from these under-explored regions using unsupervised machine learning. We will start by framing Resting-State fMRI as a massive time-series dataset, where each 3D pixel (voxel) in the brain contains a temporal signal of brain activity (movie of the 3D Brain). By calculating Functional Connectivity, essentially the Pearson correlation between these signals, we can treat the brain as a complex graph. The core of the talk will detail how we use K-means clustering on voxel-wise connectivity matrices to identify latent brain networks. We will dive into the engineering challenges of this approach, specifically how to decide the value of K. Finally, we will look at the results: how these data-driven networks allow us to identify group differences and behavioral associations, effectively turning raw, noisy 4D imagery into a structured feature set for clinical analysis. Speaker(s): Pratik Jain, Agenda: Hybrid event, in-person or online: 2:30 Introduction Technical Talks Question and Answer Networking 4:30 Conclusion Room: Meeting Rooms 2,3, 2 Civic Center Drive, East Brunswick, New Jersey, United States, 08816, Virtual: https://events.vtools.ieee.org/m/560974