Evaluation of Self-Supervised Learning for Vehicle-Type Detection in Autonomous Driving

Abstract

Self-Supervised Learning (SSL) method was implemented on vehicle-type detection task within autonomous driving context. The task involved evaluation of prominent SSL methods such as SimCLR and SimSiam against the conventional supervised method. The study explores the influence of different factors on the performance of SSL and its ability to generalize well under different circumstances.

Description

Citation

Endorsement

Review

Supplemented By

Referenced By