DeepAI AI Chat
Log In Sign Up

2nd Place Solution for Waymo Open Dataset Challenge – Real-time 2D Object Detection

by   Yueming Zhang, et al.

In an autonomous driving system, it is essential to recognize vehicles, pedestrians and cyclists from images. Besides the high accuracy of the prediction, the requirement of real-time running brings new challenges for convolutional network models. In this report, we introduce a real-time method to detect the 2D objects from images. We aggregate several popular one-stage object detectors and train the models of variety input strategies independently, to yield better performance for accurate multi-scale detection of each category, especially for small objects. For model acceleration, we leverage TensorRT to optimize the inference time of our detection pipeline. As shown in the leaderboard, our proposed detection framework ranks the 2nd place with 75.00 Waymo Open Dataset Challenges, while our framework achieves the latency of 45.8ms/frame on an Nvidia Tesla V100 GPU.


page 2

page 3


2nd Place Solution for Waymo Open Dataset Challenge – 2D Object Detection

A practical autonomous driving system urges the need to reliably and acc...

Confidence Trigger Detection: An Approach to Build Real-time Tracking-by-detection System

With deep learning based image analysis getting popular in recent years,...

CenterAtt: Fast 2-stage Center Attention Network

In this technical report, we introduce the methods of HIKVISION_LiDAR_De...

Optimizing Anchor-based Detectors for Autonomous Driving Scenes

This paper summarizes model improvements and inference-time optimization...

Real-Time Anchor-Free Single-Stage 3D Detection with IoU-Awareness

In this report, we introduce our winning solution to the Real-time 3D De...

Efficient Video Understanding via Layered Multi Frame-Rate Analysis

One of the greatest challenges in the design of a real-time perception s...

Fast and Accurate Object Detection on Asymmetrical Receptive Field

Object detection has been used in a wide range of industries. For exampl...