arXiv subjects

Jungyeon Park

Publications and source records attributed to Jungyeon Park.

5 recordsLinked to original sources

arXiv2026

Simulation of prosthetic vision with PRIMA system and enhancement of face representation

Objective. Patients implanted with the PRIMA photovoltaic subretinal prosthesis in geographic atrophy report form vision with the average acuity matching the 100um pixel size. Although this remarkable outcome enables them to read and write, they report difficulty with perceiving faces. Despite the pixelated stimulation, patients see smooth patterns rather than dots. We present a novel, non-pixelated algorithm for simulating prosthetic vision, compare its predictions to clinical outcomes, and describe computer vision and machine learning (ML) methods to improve face representation. Approach. Our simulation algorithm (ProViSim) integrates a spatial resolution filter based on sampling density limited by the pixel pitch and a contrast filter representing reduced contrast sensitivity of prosthetic vision. Patterns of Landolt C and human faces created using this simulator are compared to reports from actual PRIMA users. To recover the facial features lost in prosthetic vision due to limited resolution or contrast, we apply an ML facial landmarking model, as well as contrast-adjusting tone curves to the image prior to its projection onto the photovoltaic retinal implant. Main results. Prosthetic vision simulated using the above algorithm matches the letter acuity observed in clinical studies, as well as the patients' descriptions of perceived facial features. Applying the inversed contrast filter to images prior to projection onto the implant and accentuating the facial features using an ML facial landmarking model helps preserve the contrast in prosthetic vision, improves emotion recognition and reduces the response time. Significance. Spatial and contrast constraints of prosthetic vision limit the resolvable features and degrade natural images. ML based methods and contrast adjustments prior to image projection onto the implant mitigate some limitations and improve face representation.

cs.HC↗

arXiv2025

LVT: Large-Scale Scene Reconstruction via Local View Transformers

Large transformer models are proving to be a powerful tool for 3D vision and novel view synthesis. However, the standard Transformer's well-known quadratic complexity makes it difficult to scale these methods to large scenes. To address this challenge, we propose the Local View Transformer (LVT), a large-scale scene reconstruction and novel view synthesis architecture that circumvents the need for the quadratic attention operation. Motivated by the insight that spatially nearby views provide more useful signal about the local scene composition than distant views, our model processes all information in a local neighborhood around each view. To attend to tokens in nearby views, we leverage a novel positional encoding that conditions on the relative geometric transformation between the query and nearby views. We decode the output of our model into a 3D Gaussian Splat scene representation that includes both color and opacity view-dependence. Taken together, the Local View Transformer enables reconstruction of arbitrarily large, high-resolution scenes in a single forward pass. See our project page for results and interactive demos https://toobaimt.github.io/lvt/.

cs.CV↗

arXiv2025

Scaling Transformer-Based Novel View Synthesis Models with Token Disentanglement and Synthetic Data

Large transformer-based models have made significant progress in generalizable novel view synthesis (NVS) from sparse input views, generating novel viewpoints without the need for test-time optimization. However, these models are constrained by the limited diversity of publicly available scene datasets, making most real-world (in-the-wild) scenes out-of-distribution. To overcome this, we incorporate synthetic training data generated from diffusion models, which improves generalization across unseen domains. While synthetic data offers scalability, we identify artifacts introduced during data generation as a key bottleneck affecting reconstruction quality. To address this, we propose a token disentanglement process within the transformer architecture, enhancing feature separation and ensuring more effective learning. This refinement not only improves reconstruction quality over standard transformers but also enables scalable training with synthetic data. As a result, our method outperforms existing models on both in-dataset and cross-dataset evaluations, achieving state-of-the-art results across multiple benchmarks while significantly reducing computational costs. Project page: https://scaling3dnvs.github.io/

cs.GR↗

arXiv2023

ELIXR: Towards a general purpose X-ray artificial intelligence system through alignment of large language models and radiology vision encoders

In this work, we present an approach, which we call Embeddings for Language/Image-aligned X-Rays, or ELIXR, that leverages a language-aligned image encoder combined or grafted onto a fixed LLM, PaLM 2, to perform a broad range of chest X-ray tasks. We train this lightweight adapter architecture using images paired with corresponding free-text radiology reports from the MIMIC-CXR dataset. ELIXR achieved state-of-the-art performance on zero-shot chest X-ray (CXR) classification (mean AUC of 0.850 across 13 findings), data-efficient CXR classification (mean AUCs of 0.893 and 0.898 across five findings (atelectasis, cardiomegaly, consolidation, pleural effusion, and pulmonary edema) for 1% (~2,200 images) and 10% (~22,000 images) training data), and semantic search (0.76 normalized discounted cumulative gain (NDCG) across nineteen queries, including perfect retrieval on twelve of them). Compared to existing data-efficient methods including supervised contrastive learning (SupCon), ELIXR required two orders of magnitude less data to reach similar performance. ELIXR also showed promise on CXR vision-language tasks, demonstrating overall accuracies of 58.7% and 62.5% on visual question answering and report quality assurance tasks, respectively. These results suggest that ELIXR is a robust and versatile approach to CXR AI.

cs.CV↗

arXiv2023

Probabilistic Forecast-based Portfolio Optimization of Electricity Demand at Low Aggregation Levels

In the effort to achieve carbon neutrality through a decentralized electricity market, accurate short-term load forecasting at low aggregation levels has become increasingly crucial for various market participants' strategies. Accurate probabilistic forecasts at low aggregation levels can improve peer-to-peer energy sharing, demand response, and the operation of reliable distribution networks. However, these applications require not only probabilistic demand forecasts, which involve quantification of the forecast uncertainty, but also determining which consumers to include in the aggregation to meet electricity supply at the forecast lead time. While research papers have been proposed on the supply side, no similar research has been conducted on the demand side. This paper presents a method for creating a portfolio that optimally aggregates demand for a given energy demand, minimizing forecast inaccuracy of overall low-level aggregation. Using probabilistic load forecasts produced by either ARMA-GARCH models or kernel density estimation (KDE), we propose three approaches to creating a portfolio of residential households' demand: Forecast Validated, Seasonal Residual, and Seasonal Similarity. An evaluation of probabilistic load forecasts demonstrates that all three approaches enhance the accuracy of forecasts produced by random portfolios, with the Seasonal Residual approach for Korea and Ireland outperforming the others in terms of both accuracy and computational efficiency.

q-fin.RM↗