IMPROVING RAIL TRACKS DEFECT CLASSIFICATION BASED ON A CASCADE SWIN-TRANSFORMER MODEL

Kam Chuen Yung; Hang Shan Choy; NG Vincent To Yee; Kwan Ho Lin; XUE, Kaiwen; Kin Man Lam

IMPROVING RAIL TRACKS DEFECT CLASSIFICATION BASED ON A CASCADE SWIN-TRANSFORMER MODEL

Kam Chuen Yung (Inventor), Hang Shan Choy (Inventor), NG Vincent To Yee (Inventor), Kwan Ho Lin (Inventor), XUE, Kaiwen (Inventor), Kin Man Lam (Inventor)

Research output: Patents, agreements, assignments and companies › Patents granted

Abstract

The present invention relates to a computer-implementable system and method for detecting defects in railway tracks images using Artificial Intelligence (AI). The system comprises a receiving module configured to receive image inputs, a pre-processing module configured to pre-process the image inputs, a two-stage Transformer model configured to detect defects and false alarms based on the
image inputs, and an output module to display results containing images with defects, false alarms or unknown

Original language	English
Patent number	HK30096108
Filing date	31/10/23
Publication status	Published - 9 Feb 2024

Cite this

@misc{8cffc6d6bf454f64b92d6e628db5c92b,

title = "IMPROVING RAIL TRACKS DEFECT CLASSIFICATION BASED ON A CASCADE SWIN-TRANSFORMER MODEL",

abstract = "The present invention relates to a computer-implementable system and method for detecting defects in railway tracks images using Artificial Intelligence (AI). The system comprises a receiving module configured to receive image inputs, a pre-processing module configured to pre-process the image inputs, a two-stage Transformer model configured to detect defects and false alarms based on theimage inputs, and an output module to display results containing images with defects, false alarms or unknown",

author = "Yung, {Kam Chuen} and Choy, {Hang Shan} and {NG Vincent To Yee} and Lin, {Kwan Ho} and {XUE, Kaiwen} and Lam, {Kin Man}",

year = "2024",

month = feb,

day = "9",

language = "English",

type = "Patent",

note = "HK30096108",

}

TY - PAT

T1 - IMPROVING RAIL TRACKS DEFECT CLASSIFICATION BASED ON A CASCADE SWIN-TRANSFORMER MODEL

AU - Yung, Kam Chuen

AU - Choy, Hang Shan

AU - NG Vincent To Yee,

AU - Lin, Kwan Ho

AU - XUE, Kaiwen,

AU - Lam, Kin Man

PY - 2024/2/9

Y1 - 2024/2/9

N2 - The present invention relates to a computer-implementable system and method for detecting defects in railway tracks images using Artificial Intelligence (AI). The system comprises a receiving module configured to receive image inputs, a pre-processing module configured to pre-process the image inputs, a two-stage Transformer model configured to detect defects and false alarms based on theimage inputs, and an output module to display results containing images with defects, false alarms or unknown

AB - The present invention relates to a computer-implementable system and method for detecting defects in railway tracks images using Artificial Intelligence (AI). The system comprises a receiving module configured to receive image inputs, a pre-processing module configured to pre-process the image inputs, a two-stage Transformer model configured to detect defects and false alarms based on theimage inputs, and an output module to display results containing images with defects, false alarms or unknown

M3 - Patents granted

M1 - HK30096108

Y2 - 2023/10/31

ER -

IMPROVING RAIL TRACKS DEFECT CLASSIFICATION BASED ON A CASCADE SWIN-TRANSFORMER MODEL

Abstract

Fingerprint

Cite this