Neural Batch Sampling with Reinforcement Learning for Semi-Supervised Anomaly Detection

no code implementations • ECCV 2020 • Wen-Hsuan Chu, Kris M. Kitani

In this work, our key hypothesis is that this change in loss values during training can be used as a feature to identify anomalous data.

reinforcement-learning Reinforcement Learning (RL) +2

DreamScene4D: Dynamic Multi-Object Scene Generation from Monocular Videos

no code implementations • 3 May 2024 • Wen-Hsuan Chu, Lei Ke, Katerina Fragkiadaki

We first decompose the video scene by using open-vocabulary mask trackers and an adapted image diffusion model to segment, track, and amodally complete the objects and background in the video.

Depth Estimation Depth Prediction +3

Zero-Shot Open-Vocabulary Tracking with Large Pre-Trained Models

no code implementations • 10 Oct 2023 • Wen-Hsuan Chu, Adam W. Harley, Pavel Tokmakov, Achal Dave, Leonidas Guibas, Katerina Fragkiadaki

This begs the question: can we re-purpose these large-scale pre-trained static image models for open-vocabulary video tracking?

Object Object Tracking +5

Spot and Learn: A Maximum-Entropy Patch Sampler for Few-Shot Image Classification

no code implementations • CVPR 2019 • Wen-Hsuan Chu, Yu-Jhe Li, Jing-Cheng Chang, Yu-Chiang Frank Wang

Few-shot learning (FSL) requires one to learn from object categories with a small amount of training data (as novel classes), while the remaining categories (as base classes) contain a sufficient amount of data for training.

Data Augmentation Few-Shot Image Classification +2