Author name cluster

Lili Ju

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

1 author row

AAAI Conference 2025 Conference Paper

Frequency-Aware Density Control via Reparameterization for High-Quality Rendering of 3D Gaussian Splatting

Zhaojie Zeng
Yuesong Wang
Lili Ju
Tao Guan

By adaptively controlling the density and generating more Gaussians in regions with high-frequency information, 3D Gaussian Splatting (3DGS) can better represent scene details. From the signal processing perspective, representing details usually needs more Gaussians with relatively smaller scales. However, 3DGS currently lacks an explicit constraint linking the density and scale of 3D Gaussians across the domain, leading to 3DGS using improper-scale Gaussians to express frequency information, resulting in the loss of accuracy. In this paper, we propose to establish a direct relation between density and scale through the reparameterization of the scaling parameters and ensure the consistency between them via explicit constraints (i.e., density responds well to changes in frequency). Furthermore, we develop a frequency-aware density control strategy, consisting of densification and deletion, to improve representation quality with fewer Gaussians. A dynamic threshold encourages densification in high-frequency regions, while a scale-based filter deletes Gaussians with improper scale. Experimental results on various datasets demonstrate that our method outperforms existing state-of-the-art methods quantitatively and qualitatively.

PDF Details DOI

AAAI Conference 2022 Conference Paper

Style Mixing and Patchwise Prototypical Matching for One-Shot Unsupervised Domain Adaptive Semantic Segmentation

Xinyi Wu
Zhenyao Wu
Yuhang Lu
Lili Ju
Song Wang

In this paper, we tackle the problem of one-shot unsupervised domain adaptation (OSUDA) for semantic segmentation where the segmentors only see one unlabeled target image during training. In this case, traditional unsupervised domain adaptation models usually fail since they cannot adapt to the target domain with over-fitting to one (or few) target samples. To address this problem, existing OSUDA methods usually integrate a style-transfer module to perform domain randomization based on the unlabeled target sample, with which multiple domains around the target sample can be explored during training. However, such a style-transfer module relies on an additional set of images as style reference for pre-training and also increases the memory demand for domain adaptation. Here we propose a new OSUDA method that can effectively relieve such computational burden. Specifically, we integrate several style-mixing layers into the segmentor which play the role of style-transfer module to stylize the source images without introducing any learned parameters. Moreover, we propose a patchwise prototypical matching (PPM) method to weighted consider the importance of source pixels during the supervised training to relieve the negative adaptation. Experimental results show that our method achieves new state-of-the-art performance on two commonly used benchmarks for domain adaptive semantic segmentation under the one-shot setting and is more efficient than all comparison approaches.

PDF Details

AAAI Conference 2021 Conference Paper

Binaural Audio-Visual Localization

Xinyi Wu
Zhenyao Wu
Lili Ju
Song Wang

Localizing sound sources in a visual scene has many important applications and quite a few traditional or learning-based methods have been proposed for this task. Humans have the ability to roughly localize sound sources within or beyond the range of the vision using their binaural system. However most existing methods use monaural audio, instead of binaural audio, as a modality to help the localization. In addition, prior works usually localize sound sources in the form of object-level bounding boxes in images or videos and evaluate the localization accuracy by examining the overlap between the ground-truth and predicted bounding boxes. This is too rough since a real sound source is often only a part of an object. In this paper, we propose a deep learning method for pixel-level sound source localization by leveraging both binaural recordings and the corresponding videos. Specifically, we design a novel Binaural Audio-Visual Network (BAVNet), which concurrently extracts and integrates features from binaural recordings and videos. We also propose a point-annotation strategy to construct pixel-level ground truth for network training and performance evaluation. Experimental results on Fair-Play and YT-Music datasets demonstrate the effectiveness of the proposed method and show that binaural audio can greatly improve the performance of localizing the sound sources, especially when the quality of the visual information is limited.

PDF Details

AAAI Conference 2020 Conference Paper

SalSAC: A Video Saliency Prediction Model with Shuffled Attentions and Correlation-Based ConvLSTM

Xinyi Wu
Zhenyao Wu
Jinglin Zhang
Lili Ju
Song Wang

The performance of predicting human ﬁxations in videos has been much enhanced with the help of development of the convolutional neural networks (CNN). In this paper, we propose a novel end-to-end neural network “SalSAC” for video saliency prediction, which uses the CNN-LSTM-Attention as the basic architecture and utilizes the information from both static and dynamic aspects. To better represent the static information of each frame, we ﬁrst extract multi-level features of same size from different layers of the encoder CNN and calculate the corresponding multi-level attentions, then we randomly shufﬂe these attention maps among levels and multiply them to the extracted multi-level features respectively. Through this way, we leverage the attention consistency across different layers to improve the robustness of the network. On the dynamic aspect, we propose a correlationbased ConvLSTM to appropriately balance the inﬂuence of the current and preceding frames to the prediction. Experimental results on the DHF1K, Hollywood2 and UCF-sports datasets show that SalSAC outperforms many existing stateof-the-art methods.

PDF Details

YNIMG Journal 2005 Journal Article

Quantitative evaluation of three cortical surface flattening methods

Lili Ju
Monica K. Hurdal
Josh Stern
Kelly Rehm
Kirt Schaper
David Rottenberg

During the past decade, several computational approaches have been proposed for the task of mapping highly convoluted surfaces of the human brain to simpler geometric objects such as a sphere or a topological disc. We report the results of a quantitative comparison of FreeSurfer, CirclePack, and LSCM with respect to measurements of geometric distortion and computational speed. Our results indicate that FreeSurfer performs best with respect to a global measurement of metric distortion, whereas LSCM performs best with respect to angular distortion and best in all but one case with a local measurement of metric distortion. FreeSurfer provides more homogeneous distribution of metric distortion across the whole cortex than CirclePack and LSCM. LSCM is the most computationally efficient algorithm for generating spherical maps, while CirclePack is extremely fast for generating planar maps from patches.

Details DOI