method: GoMatching2024-04-25

Authors: HeHaibin, YeMaoyuan, ZhangJing, LiuJuhua, TaoDacheng

Affiliation: Wuhan University

Description: We extend off-the-shelf image text spotter DeepSolo to video text spotter via long-short term matching module.

GoMatching: A Simple Baseline for Video Text Spotting via Long and Short Term Matching

https://arxiv.org/abs/2401.07080

Source code

method: TencentOCR2023-03-21

Authors: Yuxin Wang, Qingxiang Lin, Sicong Liu, Huiwen Shi, Fan Yang, Lifu Wang, Haoxi Li, Weida Chen, Chen Li, Mingmin Yang, Chunchao Guo, Hongfa Wang, Wei Liu

Affiliation: TencentOCR

Description: We integrated the detection results of DBNet and Cascade MaskRCNN built with multiple Backbone architectures, combined with the Parseq English recognition model for recognition, and further improved the end-to-end tracking with ByteTrack. As a result, we obtained end-to-end tracking and trajectory recognition results.

method: DA2023-03-21

Authors: SheShuang, WangChengGuo, GaoHaiChao, DengYuCheng, DuanMengFei

Affiliation: GuangZhou

Description: Inspired by the BOT-SORT, we optimized and improved the VideoTextSCM method, called VideoTextSCM-Final.

Ranking Table

Description Paper Source Code
DateMethodMOTAMOTPIDF1Mostly MatchedPartially MatchedMostly Lost
2024-04-25GoMatching22.83%80.43%46.09%267016708237
2023-03-21TencentOCR22.44%80.82%56.45%506210756440
2023-03-21DA10.51%78.97%53.45%462913926556
2023-03-21abcmot5.54%74.61%24.25%52894611103
2023-03-15TransDETR0.00%0.00%0.00%000
2023-03-18End-to-End0.00%0.00%0.00%000
2023-03-21solar flare0.00%0.00%0.00%000
2023-03-19TextTrack-25.09%74.95%26.38%1388112710062
2023-03-19TextTrack-25.09%74.95%26.38%1388112710062
2023-03-19SCUT-MMOCR-KS-27.47%76.59%43.61%409014717016
2023-03-20TransDeTR+HRNet-28.58%80.36%26.20%155654310478

Ranking Graphic