The official PyTorch code implementation of "Personalized Trajectory Prediction via Distribution Discrimination" in ICCV 2021.

Related tags

Deep LearningDisDis
Overview

Personalized Trajectory Prediction via Distribution Discrimination (DisDis)

The official PyTorch code implementation of "Personalized Trajectory Prediction via Distribution Discrimination" in ICCV 2021,arxiv.

Introduction

The motivation of DisDis is to learn the latent distribution to represent different motion patterns, where the motion pattern of each person is personalized due to his/her habit. We learn the distribution discriminator in a self-supervised manner, which encourages the latent variable distributions of the same motion pattern to be similar while pushing the ones of the different motion patterns away. DisDis is a plug-and-play module which could be integrated with existing multi-modal stochastic predictive models to enhance the discriminative ability of latent distribution. Besides, we propose a new evaluation metric for stochastic trajectory prediction methods. We calculate the probability cumulative minimum distance (PCMD) curve to comprehensively and stably evaluate the learned model and latent distribution, which cumulatively selects the minimum distance between sampled trajectories and ground-truth trajectories from high probability to low probability. PCMD considers the predictions with corresponding probabilities and evaluates the prediction model under the whole latent distribution.

image Figure 1. Training process for the DisDis method. DisDis regards the latent distribution as the motion pattern and optimizes the trajectories with the same motion pattern to be close while the ones with different patterns are pushed away, where the same latent distributions are in the same color. For a given history trajectory, DisDis predicts a latent distribution as the motion pattern, and takes the latent distribution as the discrimination to jointly optimize the embeddings of trajectories and latent distributions.

Requirements

  • Python 3.6+
  • PyTorch 1.4

To build all the dependency, you can follow the instruction below.

pip install -r requirements.txt

Our code is based on Trajectron++. Please cite it if it's useful.

Dataset

The preprocessed data splits for the ETH and UCY datasets are in experiments/pedestrians/raw/. Before training and evaluation, execute the following to process the data. This will generate .pkl files in experiments/processed.

cd experiments/pedestrians
python process_data.py

The train/validation/test/ splits are the same as those found in Social GAN.

Model training

You can train the model for zara1 dataset as

python train.py --eval_every 10 --vis_every 1 --train_data_dict zara1_train.pkl --eval_data_dict zara1_val.pkl --offline_scene_graph yes --preprocess_workers 2 --log_dir ../experiments/pedestrians/models --log_tag _zara1_disdis --train_epochs 100 --augment --conf ../experiments/pedestrians/models/config/config_zara1.json --device cuda:0

The pre-trained models can be found in experiments/pedestrians/models/. And the model configuration is in experiments/pedestrians/models/config/.

Model evaluation

To reproduce the PCMD results in Table 1, you can use

python evaluate_pcmd.py --node_type PEDESTRIAN --data ../processed/zara1_test.pkl --model models/zara1_pretrain --checkpoint 100

To use the most-likely strategy, you can use

python evaluate_mostlikely_z.py --node_type PEDESTRIAN --data ../processed/zara1_test.pkl --model models/zara1_pretrain --checkpoint 100

Welcome to use our PCMD evaluation metric in your experiments. It is a more comprehensive and stable evaluation metric for stochastic trajectory prediction methods.

Citation

The bibtex of our paper 'Personalized Trajectory Prediction via Distribution Discrimination' is provided below:

@inproceedings{Disdis,
  title={Personalized Trajectory Prediction via Distribution Discrimination},
  author={Chen, Guangyi and Li, Junlong and Zhou, Nuoxing and Ren, Liangliang and Lu, Jiwen},
  booktitle={ICCV},
  year={2021}
}
Rainbow: Combining Improvements in Deep Reinforcement Learning

Rainbow Rainbow: Combining Improvements in Deep Reinforcement Learning [1]. Results and pretrained models can be found in the releases. DQN [2] Double

Kai Arulkumaran 1.4k Dec 29, 2022
Processed, version controlled history of Minecraft's generated data and assets

mcmeta Processed, version controlled history of Minecraft's generated data and assets Repository structure Each of the following branches has a commit

Misode 75 Dec 28, 2022
An end-to-end image translation model with weight-map for color constancy

CCUnet An end-to-end image translation model with weight-map for color constancy 1. Download the dataset (take Colorchecker_recommended dataset as an

Jianhui Qiu 1 Dec 21, 2021
E2EDNA2 - An automated pipeline for simulation of DNA aptamers complexed with small molecules and short peptides

E2EDNA2 - An automated pipeline for simulation of DNA aptamers complexed with small molecules and short peptides

11 Nov 08, 2022
ARAE-Tensorflow for Discrete Sequences (Adversarially Regularized Autoencoder)

ARAE Tensorflow Code Code for the paper Adversarially Regularized Autoencoders for Generating Discrete Structures by Zhao, Kim, Zhang, Rush and LeCun

19 Nov 12, 2021
PaddlePaddle GAN library, including lots of interesting applications like First-Order motion transfer, wav2lip, picture repair, image editing, photo2cartoon, image style transfer, and so on.

English | 简体中文 PaddleGAN PaddleGAN provides developers with high-performance implementation of classic and SOTA Generative Adversarial Networks, and s

6.4k Jan 09, 2023
A Light CNN for Deep Face Representation with Noisy Labels

A Light CNN for Deep Face Representation with Noisy Labels Citation If you use our models, please cite the following paper: @article{wulight, title=

Alfred Xiang Wu 715 Nov 05, 2022
Cours d'Algorithmique Appliquée avec Python pour BTS SIO SISR

Course: Introduction to Applied Algorithms with Python (in French) This is the source code of the website for the Applied Algorithms with Python cours

Loic Yvonnet 0 Jan 27, 2022
Intelligent Video Analytics toolkit based on different inference backends.

English | 中文 OpenIVA OpenIVA is an end-to-end intelligent video analytics development toolkit based on different inference backends, designed to help

Quantum Liu 15 Oct 27, 2022
A deep learning object detector framework written in Python for supporting Land Search and Rescue Missions.

AIR: Aerial Inspection RetinaNet for supporting Land Search and Rescue Missions AIR is a deep learning based object detection solution to automate the

Accenture 13 Dec 22, 2022
A highly efficient, fast, powerful and light-weight anime downloader and streamer for your favorite anime.

AnimDL - Download & Stream Your Favorite Anime AnimDL is an incredibly powerful tool for downloading and streaming anime. Core features Abuses the dev

KR 759 Jan 08, 2023
A machine learning benchmark of in-the-wild distribution shifts, with data loaders, evaluators, and default models.

WILDS is a benchmark of in-the-wild distribution shifts spanning diverse data modalities and applications, from tumor identification to wildlife monitoring to poverty mapping.

P-Lambda 437 Dec 30, 2022
Unofficial implementation of the Involution operation from CVPR 2021

involution_pytorch Unofficial PyTorch implementation of "Involution: Inverting the Inherence of Convolution for Visual Recognition" by Li et al. prese

Rishabh Anand 46 Dec 07, 2022
Underwater industrial application yolov5m6

This project wins the intelligent algorithm contest finalist award and stands out from over 2000teams in China Underwater Robot Professional Contest, entering the final of China Underwater Robot Prof

8 Nov 09, 2022
GAN-based Matrix Factorization for Recommender Systems

GAN-based Matrix Factorization for Recommender Systems This repository contains the datasets' splits, the source code of the experiments and their res

Ervin Dervishaj 9 Nov 06, 2022
Angora is a mutation-based fuzzer. The main goal of Angora is to increase branch coverage by solving path constraints without symbolic execution.

Angora Angora is a mutation-based coverage guided fuzzer. The main goal of Angora is to increase branch coverage by solving path constraints without s

833 Jan 07, 2023
UNION: An Unreferenced Metric for Evaluating Open-ended Story Generation

UNION Automatic Evaluation Metric described in the paper UNION: An UNreferenced MetrIc for Evaluating Open-eNded Story Generation (EMNLP 2020). Please

50 Dec 30, 2022
AugLiChem - The augmentation library for chemical systems.

AugLiChem Welcome to AugLiChem! The augmentation library for chemical systems. This package supports augmentation for both crystaline and molecular sy

BaratiLab 17 Jan 08, 2023
The devkit of the nuPlan dataset.

The devkit of the nuPlan dataset.

Motional 264 Jan 03, 2023
Creating Artificial Life with Reinforcement Learning

Although Evolutionary Algorithms have shown to result in interesting behavior, they focus on learning across generations whereas behavior could also be learned during ones lifetime.

Maarten Grootendorst 49 Dec 21, 2022