arXiv subjects

Seungjin Kim

Publications and source records attributed to Seungjin Kim.

4 recordsLinked to original sources

arXiv2026

Intelligent CCTV for Urban Design: AI-Based Analysis of Soft Infrastructure at Intersections

Artificial intelligence (AI) and computer vision are transforming transportation data collection. This study introduces an AI-enabled analytics framework leveraging existing CCTV infrastructure to evaluate the impact of soft interventions, such as temporary pedestrian refuges and curb extensions, on vehicle speed and safety. Using deep learning and perspective-based speed estimation, we evaluated driver behavior before and after interventions, with repeated post-installation monitoring in Week 1 and Week 2, in Minneapolis. Findings reveal that at unsignalized intersections, mean and 85th-percentile speeds fell by up to 18.75% and 16.56%, respectively, while pass-through traffic decreased by as much as 12.2%. Signalized intersections showed comparable reductions except one location, with mean and 85th-percentile speeds dropping by up to 20.0% and 17.19%. These results demonstrate the traffic-calming effectiveness of soft infrastructure and underscore the utility of AI-powered methods for rapid, low-cost, and evidence-based transport policy evaluation.

cs.AI↗

arXiv2026

EdgeVTP: Exploration of Latency-efficient Trajectory Prediction for Edge-based Embedded Vision Applications

Vehicle trajectory prediction is central to highway perception, but deployment on roadside edge devices necessitates bounded, deterministic end-to-end latency. We present EdgeVTP, an embedded-first trajectory predictor that combines interaction-aware graph modeling with a lightweight transformer backbone and a one-shot curve decoder. By predicting future motion as compact curve parameters (anchored at the last observed position) rather than horizon-scaled autoregressive waypoints, EdgeVTP reduces decoding overhead while producing smooth trajectories. To keep runtime predictable in crowded scenes, we explicitly bound interaction complexity via a locality graph with a hard neighbor cap. Across three highway benchmarks and two Jetson-class platforms, EdgeVTP achieves the lowest measured end-to-end latency under a protocol that includes graph construction and post-processing, while attaining state-of-the-art (SotA) prediction accuracy on two of the three datasets and competitive error on other benchmarks. Our code is available at https://github.com/SeungjinStevenKim/EdgeVTP.

cs.CV↗

arXiv2025

Dynamic VLM-Guided Negative Prompting for Diffusion Models

We propose a novel approach for dynamic negative prompting in diffusion models that leverages Vision-Language Models (VLMs) to adaptively generate negative prompts during the denoising process. Unlike traditional Negative Prompting methods that use fixed negative prompts, our method generates intermediate image predictions at specific denoising steps and queries a VLM to produce contextually appropriate negative prompts. We evaluate our approach on various benchmark datasets and demonstrate the trade-offs between negative guidance strength and text-image alignment.

cs.CV↗

arXiv2019

Pumping Dynamics of Cold-Atom Experiments in a Single Vacuum Chamber

A nonlinear analytical model for the pressure dynamics in a vacuum chamber, pumped with a sputter ion pump (SIP), is proposed, discussed and experimentally evaluated. The model describes the physics of the pumping mechanism of SIPs in the context of a cold-atom experiment. By using this model, we fit pump-down curves of our vacuum system to extract the relevant physical parameters characterizing its pressure dynamics. The aim of this investigation is the optimization of cold-atom experiments in terms of reducing the dead time for quantum sensing using atom interferometry. We develop a calibration method to improve the precision in pressure measurements via the ion current in SIPs. Our method is based on a careful analysis of the gas conductance and pumping in order to reliably link the pressure readings at the SIP with the actual pressure in the vacuum (science) chamber. Our results are in agreement with the existence of essentially two pumping regimes determined by the pressure level in the system. In particular, we find our results in agreement with the well-known fact that for a given applied voltage, at low pressures, the discharge current efficiently sputters pumping material from the pump's electrodes. This process sets the leading pumping mechanism in this limit. At high pressures, the discharge current drops and the pumping is mainly performed by the already sputtered material.

physics.atom-ph↗